Sei sulla pagina 1di 7

M4entally Disordered WVom-en in Jail:

WVho Receives Services?

...........

Linda A. Teplin, PhD, Karen M. Abram, PhD, and Gary M. McClelland, PhD

Introduction Female offenders may be more likely than


men to be defined as "mad" rather than
Many jail inmates have severe psy- "bad."233 Steadman et al.3 found women
chiatric disorders and require mental disproportionately represented among per-
health services.'-9 Numerous court deci- sons found "unfit to stand trial" or "not
sions have established that mentally ill jail guilty by reason of insanity," suggesting
inmates have a constitutional right to that the criminal justice system may be
treatment'IO'1' and that they should receive more likely to treat female offenders than
services on an emergency and regular male offenders as mentally ill.
basis.'0"21-"6 The law equates inmates' Third, the resources available in
right to mental health services with their women's correctional facilities are differ-
right to medical treatment. 12'17'18 ent from those available in men's facili-
Despite the law, mental health profes- ties. Ironically, because there are rela-
sionals speculate that many jail detainees tively few female inmates, the per capita
do not receive needed services.'19-'23 How- cost is too high to provide them with
ever, only one empirical study has exam- comparable services.33 Although the courts
ined this question. In a sample of 728 have mandated that female prisoners must
randomiy selected male jail detainees, have equivalent services34-36 and must be
Teplin found that 37% of those who had a provided adequate health care,373 in
severe mental disorder (schizophrenia or practice, they are relatively underserved.4
major affective disorder) received ser- In this study, we investigated what
vices while in jail.2 Type of diagnosis proportion of female jail detainees had
(schizophrenia vs depression) and treat- severe mental disorders and needed men-
ment history predicted who received tal health services, and, of these women,
services. what proportion received services; we
No study has yet examined whether also investigated what variables predicted
mentally ill women in jail receive needed who received services. To examine these
services. This omission is critical; the issues, we conducted structured psychiat-
female jail inmate population more than ric interviews with a random sample of
tripled between 1983 and 1994, while the female jail detainees and then collected
male population doubled.2 By 1994, the longitudinal data to find out whether those
female jail census was almost 50 000, or who needed services received them while
1 0% of the nation's jail population.2 in jail.
12 Findings from studies of men cannot
be generalized to women for three rea- Method
.....
....
....
sons. First, many studies suggest that The subjects were a stratified ran-

gender affects prevalence rates, diagnostic dom sample of 1272 female arrestees

profiles, and service utilization pattems.


For example, women in jail have higher
.....
rates of severe mental disorder,9 espe- The authors are with the Psycho-Legal Studies

Program, Northwestern University Medical


cially depression, than men in jail.6' With School, Chicago,
few exceptions,26'27 most researchers agree Requests for reprints should be sent to
that gender affects how mental disorders Linda A. Teplin, PhD, Psycho-Legal Studies

Program, Northwestern University Medical


are defined and treated.283 School, 7 10 N Lakeshore Dr, Suite 900, Chicago,
Second, men and women are man- IL 60611-3078.

aged differently in correctional settings. This paper was accepted July 29, 1996.

........

........

April 1997, Vol. 87, No. 4


Women in Jail

awaiting trial who entered Cook County ity45 46 (in contrast, see the report by
Department of Corrections (CCDOC) in Anthony et al.47). Detainees were inter- TABLE 1-Percentage of Female
Chicago, Ill, directly from pretrial arraign- viewed in private rooms in the intake area. Jail Detainees
ment between January 1991 and Septem- Most interviews lasted 1 to 3 hours, (n = 955) Who Need
ber 1993. We stratified the sample by depending on the number of reported Mental Health
arrest charge (felony or misdemeanor) symptoms. Although the DIS can be used Services and Receive
Them While in Jail
and race/ethnicity (African American, by lay interviewers, all but one of our
non-Hispanic White, Hispanic) to have interviewers had a master's degree in Receive Services
enough subjects to analyze differences psychology or clinical social work. We Need
between key subgroups. CCDOC receives used standard DIS procedures to train Services No Yes Total
approximately 70 000 admissions per year; interviewers and maintain consistency.
5500 are female.41 It is used solely for All detainees who do not immedi- No (n = 839) 89.6 10.4 100
Yes (n =116) 76.5 23.5 100
pretrial detention and for convicted misde- ately post bond and leave jail go through
meanants serving sentences of less than 1 mandatory intake procedures, including a
year. However, most detainees are in jail 2- to 3-minute mental health screen that is
because they cannot afford bail. CCDOC administered by correctional officers
is demographically similar to other large trained to administer it. Subjects flagged reflect the jail's true charge and racial/
urban jails,42 with a population composed at intake receive psychiatric services (e.g., ethnic composition. We used SUDAAN
predominantly of racial/ethnic minorities. in-depth evaluations, medications, acute 6.40 (Research Triangle Institute, Re-
All postarraignment detainees (ex- care services). Detainees who pass the search Triangle Park, NC) to correct the
cluding persons with gunshot wounds or screen can still receive services at any reported standard errors and tests of
other traumatic injuries) were eligible to time during their stay in jail. To determine significance. Additional methodological
participate, regardless of psychiatric mor- whether subjects received psychiatric and epidemiologic information is avail-
bidity, state of drug or alcohol intoxica- evaluations or other mental health ser- able elsewhere.9
tion, potential for violence, or fitness to vices while in jail, we examined jail For our first analysis, we created a
stand trial. Research interviewers selected records and case files until the subjects' fourfold table using two dichotomous
detainees during intake, using the detain- cases were disposed of by the courts or for variables, "needs services" (yes or no)
ees' unique jail identification number and 6 months, whichever came first. and "received services" (yes or no). The
a random numbers table. We paid subjects independent variable, "needs mental health
$15 for the interview. StatisticalAnalysis services," was defined on the basis of
Of 1418 detainees randomly se- established jail standards mandating that
lected, 59 (4.2%) refused to be inter- Before analyzing the substantive detainees with serious mental disorders
viewed. Another 87 women (6.1 %) agreed data pattems, we evaluated the Haw- receive treatment.23'48'49 Subjects were
to participate but did not finish the thome effect by means of a generalized defined as needing mental health services
interview for some reason (e.g., they linear model with time dependence (avail- if (1) they had schizophrenia or a major
became ill or were diverted for other able from the authors) to determine affective disorder and were symptomatic
processing). We found no pattem of whether the interviewers' presence in- within 2 weeks before the interview
selection bias; women who agreed to creased the number of jail detainees who (n = 110); (2) they had severe and defi-
participate did not differ on key demo- received services. We collected data on nite cognitive impairment and were men-
graphic variables from those who refused. the daily number of detainees who re- tally disoriented at the time of the
Design weights reflect nonresponse within ceived services for 6 months (184 days) interview (n = 3); or (3) they had a
each stratum. Our final participation rate before the study began, during the study moderate or severe substance use disorder
was 90%, and the final n was 1272. (977 days), and for 6 months after the and were mentally disoriented at the time
The subjects were between 17 and study ended (178 days). The number of of the interview (n = 3). Subjects with
67 years old; mean and median ages were detainees receiving services averaged 1.07 substance abuse disorders but no other
28 years. The stratified sample was 40.4% per day before the study began, increased mental disorder were not counted as
African-American, 33.6% non-Hispanic to 1.60 per day during the study period, needing services because inmates have no
White, 24.7% Hispanic, and 1.3% other. and dropped to 1.40 per day after the well-established constitutional right to
Nearly 80% were unemployed when they study ended. This pattem was not likely substance abuse rehabilitation, only emer-
were arrested. Mean and median educa- due to chance (P < .02) and suggests that gency detoxification treatment.23'49'50 Of
tional levels were 11 years. Nearly 80% of our presence may have sensitized jail the 955 subjects, 10.7% (unweighted
the women were mothers; 19% had four personnel to the psychiatric needs of n = 116) needed mental health services.
or more children. detainees. Subjects were defined as receiving
To determine diagnosis, we used the The analyses excluded 372 subjects services, the dependent variable, if they
National Institute of Mental Health Diag- who posted bond and left jail immediately (1) received an in-depth psychiatric evalu-
nostic Interview Schedule (DIS), Version after the interview. The remaining 955 ation, medications, or acute care services
III-R, which was developed for the Epi- subjects were in jail long enough to be (10.1%); (2) received a forensic examina-
demiologic Catchment Area program.43 screened by the jail and receive mental tion to determine fitness to stand trial
We chose the DIS because it is structured, health services. (2.2%); or (3) were sent to an outside
generates standard diagnoses44 (scored by Because we stratified our sample by psychiatric hospital (0.4%).
computer), differentiates lifetime disor- charge severity and race/ethnicity, all re- For our second analysis, we used
ders (which may be remitted) from cur- ported parameter estimates were weighted logistic regression to determine which
rent disorders, and has acceptable reliabil- by the inverse of the sampling fraction to variables-especially specific diagnoses-

April 1997, Vol. 87, No. 4 American Journal of Public Health 605
Teplin et al.

TABLE 2-Logistic Regression Predicting Whether or Not Female Jail Detainees (n = 944) Received Mental Health
Services While in Jail
Satherwaite
Crude Logit Adjusted Estimated
Odds Ratio 95% Cl Coefficient Chi-Square Odds Ratio 95% Cl

Schizophrenia/mania 9.35 3.42, 25.57 4.18 22.5*** 65.20 27.012, 157.37


Schizophrenia/mania with drug abuse or 5.22 1.31, 20.83 -4.46 12.0*** 0.01 0.003, 0.04
dependencea
Major depressive disorder 1.76 0.84, 3.68 -3.00 10.2** 0.05 0.019, 0.13
Major depressive disorder with drug 1.91 0.83, 4.43 2.53 5.4* 12.59 4.211, 37.64
abuse or dependencea
Drug abuse or dependence 1.47 0.87, 2.50 0.08 0.0 1.09 0.734,1.61
History of psychiatric treatment 56.50 27.82, 114.76 4.41 11 4.7*** 82.42 54.593,124.42
Two or more prior arrests 0.82 0.48, 1.40 -0.89 5.9* 0.41 0.285, 0.59
White race, high school education, 4.13 1.38, 12.35 2.62 8.9** 13.72 5.689, 33.10
and depressionb
Constant -2.77

Note. Four cases missing recency criteria on major depressive disorder, two cases missing recency criteria on schizophrenia/mania, and five cases
missing charge information are excluded from this analysis. All effects are entered as indicator variables. Confidence intervals and Satherwaite
chi-square statistics were computed with SUDAAN 6.40. Cl = confidence interval.
alnteraction terms are coded hierarchically; estimated odds ratios are deviations from lower-order terms.
bAnalysis of residuals and deviance statistics identified this combination of charactenstics as an important predictor of services received.
*P < .05; **P < .01; ***P < .001.

predicted who received services. For the 4. Sociodemographic characteris- significant, it explains less than 2% of the
logistic regression analysis, we counted a tics. Categorical variables included em- variance of the "received services" vari-
subject as receiving services only if she ployment status (working or not working), able. In contrast, 10.4% of the 839 sub-
received them within 1 week of arrest. We and race (African-American, non-His- jects who were not scored as needing
used 1 week as the cutoff for this analysis panic White, or Hispanic). For the continu- services did receive them. Most likely,
because mentally ill detainees who need ous variables-age, level of education, these latter subjects either were missed by
services should receive them soon after and self-reported income (including ille- the DIS, did not meet all the DSM-IH-R
arrest, and most detainees stay in jail less gal income)-we used exploratory meth- criteria for a disorder, or developed symp-
than a week. One week allows the ods to find which form of the independent toms sometime after the interview. All but
criminal justice system time to recognize variable (linear, quadratic, or categorical) two of the subjects who needed services
and respond to a severely ill detainee. best predicted the dependent variable; this
and received them were flagged at intake.
We explored four categories of inde- procedure is standard when there are no
In the analysis of variables predict-
pendent variables: specific hypotheses.5'
ing receipt of services, none of our social
1. Current disorder (symptomatic We used reference cell coding to status variables affected whether or not
within 2 weeks before the interview). assess the effect of the independent detainees received services (Table 2).
These dichotomous variables included variables on the dependent variable, re- Hence, these variables were excluded
schizophrenia or moderate or severe ceived services (yes or no). This technique from the final model. Alcohol abuse or
manic episode (combined, because there compares each observed effect with the
dependence was excluded on the same
were too few cases to analyze separately), reference group, detainees with none of
the disorders or characteristics in the final
grounds.
moderate or severe major depressive Analysis of residuals and deviance
episode, moderate or severe drug abuse or model (the constant).52 Interaction terms
are coded hierarchically. For statistical
statistics52 revealed a number of extreme
dependence, and moderate or severe outliers. When we examined these cases,
alcohol abuse or dependence. inference, we report the corrected Sather-
waite chi-square statistic for the likeli- we found that severely depressed White
2. Treatment history. This dichoto- high school graduates had far higher odds
mous variable was scored "yes" if the hood ratio test for each predictor as
entered last. Likelihood ratio statistics are of receiving services than any model
subject reported to the intake officer that predicted. An indicator variable for this
she had ever received psychiatric services more stable than asymptotic z statistics
(inpatient or outpatient) or taken psycho- when cell sizes become small.53'54 configuration of predictors is included in
active medications. (We first analyzed the final model. However, there were not
these variables separately, but they were enough high school graduates in our
Results sample to identify which variable-
so intercorrelated that we combined them.)
3. Crime. We included the severity Of the 116 subjects who needed education, race, or diagnosis-or combi-
(misdemeanor or felony) and nature (vio- mental health services, 23.5% received nation of variables affected who received
lent or nonviolent) of the charge and the them during their jail stay (Table 1), (phi services. There was one other outlier, but
number of prior arrests (none or one, two [the correlation coefficient] = 0.125, jackknife estimates excluding this case55
or more). P < .01). Although the correlation is did not differ from those reported below.

606 American Journal of Public Health April 1997, Vol. 87, No. 4
Women in Jail

Variables predicting who received performance, and the White/high school the effects of drugs or drug withdrawal. In
services were as follows: graduate/depressed indicator contributed contrast, detainees with major depressive
1. Current disorder. Subjects with less than 1.0% to the sensitivity of the episode and drug abuse or dependence
schizophrenia or manic episode were 65 model because few subjects had these were more likely to receive services than
times more likely to receive services than characteristics. those with only the major depressive
the reference group, persons with none of disorder. Drug withdrawal may heighten
the disorders or characteristics in the final Discussion the detainee's depressive symptoms, alert-
model (the constant) (P < .001). In con- ing jail personnel that the detainee needs
trast, persons with major depression were Our data suggest that less than one help.
less likely to receive services than persons quarter of female jail detainees who had Detainees with only one prior arrest
in the reference group (estimated odds severe mental disorders and needed ser- or no arrest history were more likely to
ratio [OR] = 0.05, P < .001). Co-occur- vices received them while they were in receive services than those with two or
ring drug abuse or dependence had sig- jail. This rate of service provision oc- more prior arrests. Detainees with fewer
nificant interaction effects, but in different curred despite the observed Hawthome prior arrests may be more upset and
directions, depending on the severe disor- effect (i.e., our presence probably raised symptomatic than those who have been
jail personnel's sensitivity to the detain- arrested many times. Moreover, jail per-
der. Because co-occurring drug abuse or sonnel may be more cautious when
dependence was coded hierarchically, we ees' psychiatric symptomatology and ser-
vice needs). The true rate of service screening detainees whom they have not
calculated the interaction by multiplying provision may be lower. assessed before.
the relevant estimated odds ratios. Table 2 The detainees' diagnoses determined Our findings-especially regarding
thus shows that persons with schizophre- whether or not they received services. The the influence of diagnosis and treatment
nia or mania and comorbid drug abuse or crude percentages are instructive: 47.5% history on whether or not detainees
dependence were less likely to receive of detainees with schizophrenia or manic received services-are similar to those of
services than persons who had only episode received services, compared with Teplin's prior study of men in jail.
schizophrenia or mania (65.2 X 0.01, 15.2% of detainees with depression. However, we found that, compared with
P < .001). However, persons with major Perhaps depressed detainees are over- men, significantly fewer women in jail
depression and comorbid drug abuse or looked in the chaos of the jail milieu. received needed mental health services
dependence were more likely to receive Improving services for detainees with (23.5% vs 35.5%; P < .01); for the most
services than persons with only major depression may reduce jail suicide, which part, this difference is because women in
depression; the odds increased from 0.05 currently accounts for 36% of all jail jail have rates of depression that are four
to 0.63 (0.05 x 12.59, P < .05). inmate deaths nationally.25 times higher than rates for men6'7'9 and
2. Treatment history. A history of Treatment history profoundly af- depression is often undetected in jails
psychiatric treatment increased the odds fected whether or not detainees received (analysis available from the authors).
of receiving services more than 80 times services, superseding the "true" presence However, the findings may also differ
(P < .001) compared with the reference or absence of disorder. Of detainees with because jail services have changed in the
group. schizophrenia or manic episode and a 8 years since the data on men were
3. Crime. Having two or more prior treatment history, 75% received services, collected.
arrests reduced the odds of receiving compared with 27% of detainees who had The accuracy of our data is depen-
services (OR = 0.41, P < .01) compared these disorders but no history; 3.5% of dent on the accuracy of the DIS. Unlike
with the reference group. detainees with major depressive episode physical medicine, psychiatry has no gold
4. Sociodemographic characteris- and no treatment history received ser- standard,60 and the DIS, like all diagnostic
tics. White high school graduates experi- vices. Because CCDOC admits so many assessments, is imperfect.43 We tried to
encing a major depressive episode had detainees every day, intake officers prob- maximize the accuracy of our data by
almost 15 times the odds of receiving ably use treatment history to determine using experienced interviewers and to
services compared with the reference current service need. Treatment history is maximize specificity by defining the
group (P < .01). a useful indicator because severe mental "needs services" category so narrowly-
To assess the impact of each cat- disorders tend to persist58'59 and because including only persons who had current
egory of variables on whether or not detainees with a history of severe mental and severe psychiatric disorders. Neverthe-
detainees received services, we computed disorder could relapse under the stress of less, given the limitations of the DIS, our
the variables' contributions to the sensitiv- incarceration. However, if intake person- investigation should be replicated as
ity (true positive fraction or positive nel use treatment history as a surrogate psychiatric assessments improve.
predictive value) of the model.5657 The measure of psychiatric disorder, they will Another study limitation is that we
positive predictive value for the model is miss many ill detainees who have never did not measure whether detainees' desire
65.5%. This analysis confirmed the impor- been treated. for treatment affected whether or not they
tance of treatment history in predicting Co-occurring drug use or depen- received services. It is possible that some
whether or not jail detainees receive dence affected the probability of receiving subjects may have hidden their symptoms
mental health services. The five mental services in different ways, depending on from jail personnel because they did not
disorder and drug abuse variables contrib- the severe disorder. Detainees with schizo- wish to receive services.
uted only 2.4% to the positive predictive phrenia or manic episode and a co- The rate of service provision that we
power of the model; in contrast, history of occurring drug use disorder were less observed in CCDOC is probably much
psychiatric treatment contributed 30.1%. likely to receive services than those with better than in most jails. Unlike many
Two or more prior arrests did not contrib- only the severe disorder. Jail personnel jails,6' CCDOC screens all detainees at
ute to the model's positive predictive may have attributed these symptoms to intake for mental health problems.

April 1997, Vol. 87, No. 4 American Journal of Public Health 607
Teplin et al.

CCDOC's integrated psychiatric service and the services that they receive. How- logic Catchment Area program. Am J
system is used as a model nationwide.62 ever, many persons in the general popula- Public Health. 1990;80:663-669.
However, our data suggest that even in tion also do not receive needed mental 7. Teplin LA. Psychiatric and substance
abuse disorders among male urban jail
jails with sophisticated programs, it is health services.68'69 Some of these persons detainees. Am J Public Health. 1994;84:
difficult to detect service needs when so may not have access to services. Others 290-293.
many detainees enter overcrowded and choose to live in the community without 8. Valdiserri EV, Carroll KR, Hartl AJ. A
burdened facilities. treatment and can do so unless they are study of offenses committed by psychotic
Our findings suggest that jails may found to be dangerous to themselves or inmates in a county jail. Hosp Community
Psychiatry. 1986;37: 163-166.
need better intake assessments tailored to others.70 However, the public health sys- 9. Teplin LA, Abram KM, McClelland GM.
the needs of incarcerated women. The tem has an obligation to jail detainees The prevalence of psychiatric disorders
Brief Psychiatric Rating Scale is useful in because they are captive and, under the among incarcerated women, I: pretrial jail
jails that employ mental health profession- 14th amendment, have a constitutional detainees. Arch Gen Psychiatry, 1996;53:
als.63 The Referral Decision Scale64 has right to needed mental health services.49 505-512.
10. Estelle v Gamble, 429 US 97 (1976).
acceptable reliability and validity63 and Correctional health care is a growing 11. Bell v Wolfish, 441 US 535, n. 16, 545
can be used by lay interviewers in jails. national public health problem.71 Correc- (1979).
Neither instrument, however, has been tional populations, especially in jails, have 12. Bowring v Godwin, 551 F2d 44 (4th Cir
validated on female inmates. increased dramatically. Between 1983 and 1977).
Our data also suggest that correc- 1994, the national jail census increased 13. Inmates ofAllegheny County Jail v Pierce,
487 F Supp 638, 642-3 (WD Pa 1980).
tional officers need more training to from 223 551 to nearly 500 000.25 Al- 14. Finney v Hutto, 410 F Supp 251, 259 (ED
identify serious psychiatric disorders, es- though health services are a large compo- Ark 1976).
pecially depression, among detainees nent of correctional facilities' operating 15. Jones v Wittenberg, 323 F Supp 793 (ND
whose disorders are missed at intake. We budgets, funding is not keeping pace with Ohio 1971).
need.7' We must continue to improve 16. Finney v Mabry, 534 F Supp 1026, 1037
found that only two of the detainees who (ED Ark 1982).
needed services and yet passed the jail's screening techniques, as well as provide 17. Ramos v Lamm, 639 F2d 559, 577 (1980).
intake screen were later provided services. liaisons between the criminal justice and 18. Hoptowit v Ray, 682 F2d 1237, 1253
Correctional officers may play a major mental health systems,72 so that mentally (1982).
role in service referral, but they need ill inmates who need services receive 19. Morgan C. Developing mental health
services for local jails. Criminal Justice
training to be effective liaisons. them. O] Behav. 1981;8:259-273.
Improving mental health services in 20. Singer RG. Providing mental health ser-
jails is only the first step. Because only vices for jail inmates: legal perspectives. J
one third of jail detainees stay longer than Acknowledgments 21.
Prison Health. 1981;1:105-129.
Whitmer GE. From hospitals to jails: the
4 days,65 the public health system must This work was supported by grants R01-
fate of California's deinstitutionalized men-
also provide mental health services in the MH45583 and RO1-MH47994 from the Na-
tional Institute of Mental Health. tally ill. Am J Orthopsychiatry. 1980;50:65-
community for released detainees. How- We greatly appreciate the cooperation of 75.
ever, providing effective programs will be Cook County Sheriff Michael F. Sheahan; 22. Steadman HJ, McCarty DW, Morrissey JP.
difficult for several reasons. First, many Michael Mahoney, executive director of the The Mentally Ill in Jail: Planning for
female jail detainees with severe mental John Howard Association; and the former Cook Essential Services. New York, NY: Guil-
County Department of Corrections executive ford Press; 1989.
disorders may also have substance abuse director, J.W. Fairman, Jr. We also want to 23. Steadman HJ, ed. Effectively Addressing
or dependence.66 Comorbidity is difficult thank Cermak Health Services, especially Dr the Mental Health Needs of Jail Detainees.
to treat, and there are not enough commu- Carl Alaimo and the intake staff, as well as our Seattle, Wash: National Coalition for the
nity placements.66 Service providers often superb interviewers. Finally, we thank Chris- Mentally Ill in the Criminal Justice Sys-
topher Winship, PhD, of Harvard University tem; 1992.
view such persons as undesirable cli- 24. Teplin LA. Detecting disorder: the treat-
ents.66 and Kiang Liu, PhD, of Northwestern Univer-
sity for their astute statistical expertise and ment of mental illness among jail detain-
Second, most released detainees are Laura Coats for her tireless editing. ees. J Consult Clin Psychol. 1990;58:233-
poor, making it difficult for them to seek 236.
services and to maintain a treatment plan. 25. Jails and Jail Inmates 1993-94. Washing-
Poverty also causes many psychological References ton, DC: US Dept of Justice, Bureau of
1. Guy E, Platt JJ, Zwerling I, Bullock S. Justice Statistics; 1995. Publication NCJ-
stressors; poor persons are more likely to Mental health status of prisoners in an 151651.
be victims of crime,67 to have inadequate urban jail. Criminal Justice Behav. 1985; 12: 26. Leaf PJ, Bruce ML. Gender differences in
housing, and to live in dangerous neighbor- 29-53. the use of mental health-related services: a
hoods.29 2. Lamb HR, Grant RW. The mentally ill in re-examination. J Health Soc Behav. 1987;
an urban county jail. Arch Gen Psychiatry. 28:171-183.
Finally, motherhood complicates ser- 1982;39: 17-22. 27. Cafferata GL, Meyers SM. Pathways to
vice utilization. Over 60% of our subjects 3. Lamb R, Grant RW. Mentally ill women in psychotropic drugs. Med Care. 1990;28:
had children under 5 years of age, and a county jail. Arch Gen Psychiatry 1983;40: 285-300.
most of the sample were single parents. 363-368. 28. Hare-Mustin RT. An appraisal of the
Few mental health programs provide child 4. Monahan J, McDonough LB. Delivering relationship between women and psycho-
community mental health services to a therapy. Am Psychol. May 1983:593-601.
care. In short, successful public health county jail population: a research note. Bull 29. Russo NF. Overview: forging research
delivery for released jail detainees must Am Acad Psychiatry Law. 1980;8:28-32. priorities for women's mental health. Am
go beyond addressing psychopathology 5. Teplin LA. The criminalization of the Psychol. 1990;45:368-373.
alone; it requires a systematic network of mentally ill: speculation in search of data. 30. Loring M, Powell B. Gender, race and
Psychol Bull. 1983;94:54-67. DSM-III: a study of the objectivity of
resources. 6. Teplin LA. The prevalence of severe psychiatric diagnostic behavior. J Health
Our findings highlight the discrep- mental disorder among male urban jail Soc Behav. 1988;29: 1-22.
ancy between jail detainees' service needs detainees: comparison with the Epidemio- 31. Visher C. Gender, police arrest decisions,

608 American Journal of Public Health April 1997, Vol. 87, No. 4
Women in Jail

and notions of chivalry. Criminology. Interview Schedule and a standardized The Abuse of Jails as Mental Hospitals.
1983;21 :5-28. psychiatric diagnosis. Arch Gen Psychia- Washington, DC: Public Citizen's Health
32. Steadman HJ, Monahan J, Hartstone E, try. 1985;42:667-675. Research Group and the National Alliance
Davis SK, Robins PC. Mentally disordered 48. Cody v Hillard, 599 F Supp 1025, 1041- for the Mentally Ill; 1992.
offenders: a national survey of patients and 1044 DSD 1984. 63. Hart SD, Roesch R, Corrado RR, Cox DN.
facilities. Law Human Behav. 1982;6:31- 49. Cohen F. Legal Issues and the Mentally The Referral Decision Scale: a validation
38. Disordered Prisoner Washington, DC: study. Law Human Behav. 1993;17:611-
33. Feinman C. Women in the Criminal Justice National Institute of Corrections; 1988. 623.
System. 3rd ed. Westport, Conn: Praeger; 50. Mushlin MB. Rights of Prisoners. 2nd ed. 64. Teplin LA, Swartz J. Screening for severe
1994. Colorado Springs, Colo: Shepard's/Mc- mental disorder in jails: the development of
34. Barefield v Leach, 10282 (Dist Ct NM Graw-Hill Inc; 1993:166-167. the Referral Decision Scale. Law Human
1974). 51. Leamer EE. Specification Searches: Ad Behav. 1989;13:1-18.
35. Cooper v Morin, 91 Misc 2d 302,398 NYS Hoc Inference with Nonexperimental Data. 65. US Department of Justice. Census ofLocal
2d 36 (Monroe Co Sup Ct 1977). New York, NY: John Wiley & Sons Inc; Jails 1988. Washington, DC: Bureau of
36. Glover v Johnson, 478 F Supp 1075 (ED 1978. Justice Statistics; 1990. Publication NCJ-
Michigan 1979). 52. Hosmer DW Jr, Lemeshow S. Applied 121101.
37. Newman v Alabama, 349 F Supp 278 (MD Logistic Regression. New York, NY: John 66. Abram KM, Teplin LA. Co-occurring
Ala 1972). Wiley & Sons Inc; 1989. disorders among mentally ill jail detainees.
38. Morales v Turman, 383 F Supp 53 (ED Tex 53. McCullagh P, Nelder JA. Generalized Am Psychol. 1991;46:1036-1045.
1974). Linear Models. 2nd ed. London: Chapman 67. Belle D. Poverty and women's mental
39. Todaro v Ward, 431 F Supp 1129 (SD NY), & Hall; 1989;323-352. health. Am Psychol. 1990;45:385-389.
affd, 565 F2d 48 (2d Cir 1977). 54. Agresti A. Categorical Data Analysis. 68. Bourdon KH, Rae DS, Narrow WE,
40. Velimesis ML. Sex roles and mental health New York, NY: John Wiley & Sons Inc; Manderscheid RW, Regier DA. National
of women in prison. Professional Psychol. 1990:419-437. prevalence and treatment of mental and
198 1; 12:128-135. 55. Belsley DA, Kuh E, Welsch RE. Regres- addictive disorders. In: Manderscheid RW,
41. Zimmerman J, Keilitz I, Fitch W, Marvell sion Diagnostics: Identifying Influential Sonnenschein MA, eds. Mental Health,
TB, Holmstrup, ME. Forensic mental Data and Sources of Collinearity. New United States, 1994. Washington, DC:
health screening and evaluation in jails. York, NY: John Wiley & Sons Inc; Department of Health and Human Ser-
Williamsburg, Va: National Center for 1980:192-261. vices; 1994:22-51. DHHS publication
State Courts; 1981. Occasional Paper no. 3. 56. Metz CE. Basic principles of ROC analy- SMA 94-3000.
42. Drug Use Forecasting Annual Report. sis. Semin Nucl Med. 1978;8:283-298. 69. Kessler RC, McGonagle KA, Zhao S, et al.
Washington, DC: US Dept of Justice, 57. Hadom DC, Keeler EB, Rogers WH, Lifetime and 12-month prevalence of
National Institute of Justice; 1990. Brook RH. Assessing the Performance of DSM-III-R psychiatric disorders in the
43. Robins LN, Helzer JE, Croughan J, Ratcliff Mortality Prediction Models. Santa Monica, United States: results from the National
K. National Institute of Mental Health Calif: RAND; 1993. Comorbidity Survey. Arch Gen Psychiatry.
Diagnostic Interview Schedule: its history, 58. Kiesler CA. Public and professional myths 1994;51 :8-19.
characteristics and validity. Arch Gen about mental hospitalization: an empirical 70. Brakel SJ, Parry J, Weiner BA. The
Psychiatrn' 1981 ;38:381-389. reassessment of policy-related beliefs. Am Mentally Disabled and the Law. 3rd ed.
44. Diagnostic and Statistical Manual of Men- Psychol. 1982;37:1323-1339. Chicago, Ill: American Bar Foundation;
tal Disorders. 3rd ed. rev. Washington, DC: 59. Taube CA, Barrett SA, eds. Mental Health, 1985.
American Psychiatric Association; 1987. United States 1985. Washington, DC: US 71. American College of Physicians, National
45. Helzer JE, Robins LN, McEvoy LT, et al. A Dept of Health and Human Services; 1985. Commission on Correctional Health Care,
comparison of clinical and Diagnostic DHHS publication ADM 85-1378. American Correctional Health Services
Interview Schedule diagnoses. Arch Gen 60. Robins LN. Epidemiology: reflections on Association. The crisis in correctional
Psychiatry. 1985;42:657-666. testing the validity of psychiatric inter- health care: the impact of the National
46. Burke JD. Diagnostic categorization by the views. Arch Gen Psychiatry. 1985;42:918- Drug Control Strategy on correctional
Diagnostic Interview Schedule (DIS): a 924. health services. Ann Internal Med. 1992;
comparison with other methods of assess- 61. McCarty D, Steadman HJ, Morrissey JP. 117:71-77.
ment. In: Barrett J, Rose R, eds. Mental Issues in planning jail mental health 72. Steadman HJ, Morris SM, Dennis DL. The
Disorders in the Communit)r New York, services. Federal Probation. December diversion of mentally ill persons from jails
NY: Guilford Press; 1986:255-285. 1982: 56-63. to community-based services: a profile of
47. Anthony JC, Folstein M, Romanoski AJ, et 62. Torrey EF, Stieber J, Ezekiel J, et al. programs. Am J Public Health. 1995;85:
al. Comparison of the lay Diagnostic Criminalizing the Seriously Mentally 'll: 1630-1635.

April 1997, Vol. 87, No. 4 American Journal of Public Health 609
This article has been cited by:

1. Linda A. Teplin, Karen M. Abram, Gary M. McClelland, Jason J. Washburn, Ann K. Pikus. 2005. Detecting Mental
Disorder in Juvenile Detainees: Who Receives Services. American Journal of Public Health 95:10, 1773-1780. [Abstract]
[Full Text] [PDF] [PDF Plus]
2. Gary Michael McClelland, Linda A. Teplin, Karen M. Abram, Naomi Jacobs. 2002. HIV and AIDS Risk Behaviors
Among Female Jail Detainees: Implications for Public Health Policy. American Journal of Public Health 92:5, 818-825.
[Abstract] [Full Text] [PDF] [PDF Plus]