The prevalence of compulsive hoarding is not known, due to an absence of epidemiologic research. However, some preliminary estimates are possible. The lifetime prevalence of OCD in the U.S. population has been estimated at 1.6% (Kessler et al., 2005
). Within samples of OCD patients, 38% report at least some clinically relevant symptoms of hoarding (Samuels et al., 2002
). These data alone would suggest a prevalence of 0.6%, or approximately 1.8 million people in the United States. However, research from hoarding (as opposed to OCD) samples suggests that as many as 83% of people with significant hoarding problems do not have OCD (Frost et al., 2006
), and therefore would have been unlikely to have presented to the OCD clinics in the aforementioned studies. Thus, 1.8 million is likely a serious underestimate. Although definitive estimates are not possible, it seems clear that compulsive hoarding is highly prevalent in the population.
The present data suggest that hoarding is associated with marked occupational and other role impairment. Prior research with smaller samples suggests that hoarding is associated with greater self-reported work, social, and family impairment than are OCD and other anxiety disorders (Frost, Steketee, Williams et al., 2000
). The present study extends these findings in a much larger sample (the largest to date on this topic) by showing that, among people who meet relatively strict diagnostic criteria for hoarding, 7% are on disability and 6% report having been fired from jobs due to hoarding. Even hoarding participants who did not meet full diagnostic criteria reported an elevated rate of job loss due to hoarding, suggesting that even less severe hoarding symptoms might be associated with substantial impairment. Among currently employed hoarding participants, the majority reported that they had difficulty finding items at work, and that their workspace was taken up by clutter. Over one-fifth failed to file income tax in at least 1 of the past 5 years, and 38% brought in personal income below the current poverty threshold for a single individual.
Perhaps most striking are the psychiatric work impairment days, likely a more precise estimate of impairment than self-reported disability used in prior research. Hoarding participants reported significantly more psychiatric work impairment days than did NCS participants with all anxiety and depressive disorders; only NCS patients with bipolar disorder and non-affective psychosis (e.g., schizophrenia) reported work and role impairment to the same extent as did hoarding participants in the present sample. However, differences in sampling methods between the NCS and the present study must be taken into account. The NCS was a stratified random survey of U.S. households, whereas the present study used a convenience sample of individuals who had learned about the research via the media (e.g., television, internet). It is certainly possible that the present sampling method was biased toward more severe cases, and that milder cases might not have self-identified as having hoarding problems or would not have been motivated to seek more information or to participate in the study (alternatively, due to poor insight, very severe cases also might not have self-identified as having hoarding problems). Nevertheless, the magnitude of the differences suggests that hoarding is associated with profound role impairment.
It is also worth noting that the mean age of women in the NCS was approximately 34 years, about 15 years younger than the hoarding participants in the present sample. The maximum age of NCS women was 61, not a match to the older hoarding participants in the present sample. Age-related medical factors likely play some role in occupational and role impairment, although this concern is alleviated somewhat by the facts that: (1) participants specifically identified the number of impairment days due to psychiatric rather than medical factors; (2) the prevalence of psychiatric disorders appears to decline with age across the population (Kessler et al., 1994
); (3) in the present sample, age was unrelated to hoarding severity or role impairment days; and (4) in the present sample, hoarding severity was significantly related to impairment even when controlling for non-psychiatric medical conditions. Therefore, it seems unlikely that age accounts for the differences in impairment between the present sample and women in the NCS.
Another potential limitation of the present study is the exclusive reliance on self-report for the hoarding participants. Although several studies now show that data collection over the internet yields results comparable to those using a paper-and-pencil format (Carlbring et al., 2007
; Coles et al., 2007
; Gosling et al., 2004
), a clinical interview might be more reliable and valid than are self-report measures of hoarding, particularly given the poor insight evident in many hoarding participants (e.g., Fitch et al., 2007
In the absence of epidemiologic research, the gender and race/ethnicity distribution of compulsive hoarding is not known, although participants in clinical trials for hoarding have been primarily female (e.g., Tolin, Frost et al., 2007b
). It is not clear whether the mainly White and female sample in the present study accurately reflects the demographics of the hoarding population. Similarly, although the large number of older adults in the present study is consistent with anecdotal estimates suggesting that hoarding severity may increase with age (Steketee et al., 2001
), it remains unclear how the present sample compares to the hoarding population in terms of age. Additional research is needed to define the demographic characteristics of individuals who hoard.
It is also unclear the extent to which the reported impairments are caused by hoarding vs. other comorbid conditions. The present study did not include a thorough evaluation of comorbid psychiatric disorders, and research suggests that comorbidity in hoarding is the rule, rather than the exception (Frost et al., 2006
). Therefore, it is entirely possible that some of the impairment among individuals who hoard is associated with other conditions such as major depressive disorder, social phobia, generalized anxiety disorder, or attention deficit-hyperactivity disorder, conditions that appear strongly related to hoarding (Frost et al., 2006
; Hartl et al., 2005
; Meunier, Tolin, Frost, Steketee, Brady et al., 2006
) and to role impairment (Adler et al., 2006
; Biederman and Faraone, 2006
; Kessler, 2003
; Wittchen, 2002
). The fact that much of the variance in role impairment remains unexplained after accounting for age, gender, medical conditions, and hoarding severity is consistent with the presence of other, unknown contributors.
Although medical service utilization was not formally measured in the present study, non-psychiatric medical expenses among individuals who hoard are likely high. A large majority (78%) was either overweight or obese according to current standards, and a majority (64%) reported at least one chronic and severe medical condition—markedly greater than the rates reported by women in the general population (as reflected by the NCS sample). Patients meeting full diagnostic criteria for hoarding showed higher rates than did NCS women of all assessed medical conditions. Fibromyalgia and chronic fatigue syndrome were among the more commonly-reported conditions, and although these conditions were not assessed in the NCS, the rates in the present hoarding participants (11% and 12%, respectively) seem to exceed women's rates of these conditions in other epidemiologic research (4.9% and 0.2%, respectively) (Jason et al., 1999
; White et al., 1999
). Forty-five percent had received mental health services in the past year, a fivefold increase over women in the general population. Among those receiving such services, the amount of utilization was high (8 psychopharmacology visits, 19 outpatient therapy visits, and a remarkable 13 inpatient hospitalization days). It is noteworthy that 14% of hoarding participants report that their medical expenses were paid by public aid programs, suggesting that the cost to society is high.
Reports of external interventions differed between self-identified hoarding participants and family informants. It is possible that the family members who participated in the present study were reporting on more severe cases. That is, family members of milder cases of hoarding may have less interest or motivation to participate in such a study. According to the family informants, as many in 1 of 25 of their hoarding relatives had a child, elder, or pet removed from the home due to hoarding, and as many as 1 in 8 had been evicted or threatened with eviction. The rates of such interventions reported by the hoarding participants themselves were lower, but nevertheless striking in underscoring the severity of hoarding. Such data are consistent with the high rate of community agency involvement in cases of hoarding (Frost, Steketee, and Williams, 2000
) and the associated high cost of such involvement.
The present data do not permit estimation of the dollar amount of the economic burden of hoarding, as has been reported for other disorders (Greenberg et al., 2003
; Greenberg et al., 1999
; E. Q. Wu et al., 2005
). However, the available data suggest that the impact of compulsive hoarding on a per-person basis exceeds those of many psychiatric disorders. High costs appear likely to affect not only individuals, but also society as a whole in terms of lost work productivity, mental health service utilization, non-psychiatric medical costs, and community agency involvement. Evaluation of the overall cost and impact of compulsive hoarding in the population awaits epidemiologic research.