Available via license: CC BY-NC-ND 4.0
Content may be subject to copyright.
322 Suicide and Life-Threatening Behavior 37(3) June 2007
2007 The American Association of Suicidology
An Evaluation of Crisis Hotline Outcomes
Part 1: Nonsuicidal Crisis Callers
John Kalafat, PhD, Madelyn S. Gould, PhD, MPH,
Jimmie Lou Harris Munfakh, BA, and Marjorie Kleinman, MS
The effectiveness of telephone crisis services/hotlines, examining proximal
outcomes as measured by changes in callers’ crisis state from the beginning to the
end of their calls to eight centers in the U.S. and intermediate outcomes within 3
weeks of their calls, was evaluated. Between March 2003 and July 2004, 1,617 crisis
callers were assessed during their calls and 801 (49.5%) participated in the follow-
up assessment. Significant decreases in callers’ crisis states and hopelessness were
found during the course of the telephone session, with continuing decreases in
crisis states and hopelessness in the following weeks. A majority of callers were
provided with referrals and/or plans of actions for their concerns and approxi-
mately one third of those provided with mental health referrals had followed up
with the referral by the time of the follow-up assessment. While crisis service staff
coded these callers as nonsuicidal, at follow-up nearly 12% of them reported hav-
ing suicidal thoughts either during or since their call to the center. The need to
conduct suicide risk assessments with crisis callers and to identify strategies to
improve referral follow-up is highlighted.
Telephone crisis services (TCS) have been many maladaptive behaviors such as alcohol
use, interpersonal violence, or suicidal behav-providing crisis intervention and referral ser-
vices in the United States for over 40 years. ior involve maladaptive responses to crises;
(3) crises are characterized by increases inThe conceptual bases for crisis intervention
include: (1) crises are time limited and pre- anxiety, which produce cognitive constriction
and attenuate problem-solving ability; andsent an opportunity for positive or negative
outcomes, based on the application of effec- (4) due to the failure of the usual coping
mechanisms and heightened vigilance, indi-tive or maladaptive coping, respectively; (2)
Dr. Kalafat is a faculty member of the Rutgers Graduate School of Applied and Professional
Psychology; Dr. Gould is a Professor at Columbia University in the Division of Child and Adolescent
Psychiatry (College of Physicians & Surgeons) and Department of Epidemiology (School of Public
Health), and a Research Scientist at the New York State Psychiatric Institute; and Ms. Munfakh and
Ms. Kleinman are in the Division of Child and Adolescent Psychiatry and New York State Psychiatric
Institute.
Supported by a cooperative agreement from the Substance Abuse and Mental Health Services
Administra tion (SAMHSA), U79SM54128. We gratefully acknowledge Dr. Rachel Strohl and Ms. Rebe-
cca Brent Weinberg for their assistance in data collection and analysis, respectively. While they must
remain anonymous, we want to thank the participating centers for their dedication and hard work on the
project.
Address correspondence to John Kalafat, PhD, Rutgers Graduate School of Applied and Pro-
fessional Psychology, 152 Frelinghuysen Rd., Piscataway, NJ 08854-8085; E-mail: kalafat@rci.
rutgers.edu
Kalafat et al. 323
viduals are more open to intervention (Caplan, maladaptive coping, and to help the caller
find a plan for coping with the situation that1964; Rapoport, 1965). The implications of
these concepts for intervention suggest that precipitated the crisis and/or another helping
agency that can provide further assistance.interventions must be readily accessible to
provide adaptive responses to crises to atten- Evaluations of TCS have included
caller feedback/satisfaction, assessments ofuate maladaptive outcomes; due to the reduc-
tion in defenses, a relatively brief interven- helping processes and proximal outcomes
(changes in caller crisis or suicidal status),tion may have a significant impact; and a
collaborative intervention which includes ac- rates of follow-up with referrals, and assess-
ments of distal outcom es consisting of changestive problem solving and mobilization of in-
ternal and external resources is necessary to in community suicide rates. The focus of this
study is outcomes for nonsuicidal crisis call-take advantage of the opportunity for growth
presented by crises (Baldwin, 1979; Brock- ers to TCS. Therefore, studies of crisis and
referral outcomes for these callers are re-opp, 1973).
Based on this rationale, a model of cri- viewed. A companion article addresses TCS
responses to suicidal callers (Gould, Kalafat,sis services has evolved that consists of 24-
hour telephone services (often supplemented Harris Munfakh, & Kleinman, this issue).
Early process evaluations of telephoneby mobile outreach teams) staffed by spe-
cially trained professionals and/or parapro- counseling interventions focused on helper-
offered conditions of empathy, warmth, andfessional volunteers who provide one-time or
time-limited interventions to clients at no genuineness (Rogers, 1957), which are repre-
sentative of helping approaches of many tele-charge. A consensus has evolved around a
four to six step problem-solving intervention phone crisis centers to this day. Studies found
moderate levels of these conditions as ratedmodel first adopted by the Los Angeles Sui-
cide Prevention Center (Farberow, Heilig, & on simulated calls and role-plays, variations
between centers, and increased levels associ-Litman, 1968), consisting of establishing rap-
port; defining the problem(s) (including as- ated with training and experience (Bleach &
Claiborn, 1974; Caruthers & Inslee, 1974;sessing risk for suicide); exploring affect (in-
cluding reducing anxiety and other affects France, 1975; Kalafat, Boroto, & France,
1979; O’Donnell & George, 1977); however,that attenuate problem solving); exploring
callers’ coping repertoires; and developing al- the relationship between these conditions
and call outcomes was not assessed. Otherternatives for addressing the problem (i.e., a
specific plan of action and/or referral to in- studies have assessed the presence and timing
of the components of the helping model andformal or formal resources). In addition to a
variety of crisis situations, TCS provide im- examined their relationships to caller out-
comes through follow-up calls to callersmediate responses to suicidal callers. For
callers in less acute suicidal states who may (Echterling & Hartsough, 1989; Echterling,
Hartsough, & Zarle, 1980; Young, 1989).be at the ideation or planning stage, TCS
aim to identify the precipitants of the suicidal The presence and timing of these compo-
nents were related to positive caller feedbackstate, generate alternative coping strategies,
and mobilize supports for callers. For callers and outcomes such as relief of distress, confi-
dence, and emotional awareness.in more acute, imminent suicidal states, TCS
may engage in more active interventions such Asessments of callers’ follow through
with referral recommendations have alsoas obtaining the location of callers through
direct request, tracing calls, or employing been conducted (Buchta, Wetzel, Reich, But-
ler, & Fuller, 1973; Paul & Turner, 1976;caller identification; and, sending community
emergency response personnel. The goal of Slaikeu, Tulkin, & Speer, 1975; Slaikeu &
Willis, 1978). In general, studies found fol-telephone crisis intervention, then, is to re-
duce maladaptive cognitive and affective low through rates of approximately 50%.
Two early studies that rated the appropriate-components of the crisis state, to attenuate
324 Nonsuicidal Crisis Caller Outcomes
ness of referrals provided by telephone staff cesses, use of internal call monitoring, and
willingness to adopt agreed upon standard-raised concerns about the accuracy of refer-
rals (Bleach & Claiborn, 1974) and their ap- ization of call record keeping and evaluation
procedures. An additional three centers hadpropriateness for the caller’s problem (Apsler
& Hodas, 1975). Whether or not callers fol- been originally recruited, but dropped out of
the study because one had assumed a newlow through with referrals may also be a
function of such variables as caller motivation service and could not take on additional proj-
ects; another conducted face-to-face outreachand availability of other sources of support
(Slaikeu et al., 1975; Slaikeu & Willis, 1978). with all callers, and thus was not a typical
telephone hotline; and the staff of the thirdAlso, it may be that effective efforts to reduce
callers’ anxiety or resolve their concerns may refused to collect evaluation data. Seven of
the participating centers were members of aattenuate motivation to contact referrals. Fi-
nally, counselor skill may affect referral fol- national 1-800-SUICIDE network. The eight
centers were located in six states (2 midwest,low through; for example, Slaikeu et al.
(1975) found that callers who followed through 4 northeast, 1 south, and 1 west). Telephone
counselors in these centers were either paidrated the counselor who provided the referral
as more helpful than those who did not. In (4), volunteer (3), or a mixture of both paid
and volunteer (1). The annual call volume inaddition, Paul and Turner (1976) found that
provision of incentives for crisis staff’s con- these centers ranged from 7,993 to 85,000
calls per year. The method used by crisis cen-scientious provision of referrals increased
rates of follow through. ters to select which counselors would partici-
pate in the project varied. Three crisis cen-To date, there are no studies that eval-
uate a primary aim of crisis services—to re- ters required all of their counselors to
participate, and in two other centers counse-duce the crisis states of callers from their
presenting levels at the start of the call—and lors participated on a volunteer basis. The re-
maining three centers used specific criteria toto provide a specific plan of action or referral
for the caller’s presenting problem. The pur- select only a portion of their counselors: in
one, all but the overnight staff were requiredpose of our study was to evaluate the effec-
tiveness of TCS for achieving positive client to participate; in the second, only paid staff
was selected to participate; and in the thirdoutcomes as indicated by changes in callers’
crisis states from the beginning to the end of center only those with extensive crisis coun-
seling experience were selected. A total oftheir calls to participating centers (immediate
outcomes), and again within 3 weeks of their 240 counselors in the eight centers con-
ducted the baseline assessment with callers.calls (intermediate outcomes). Callers’ recol-
lection of and follow through with plans of The number of counselors participating in
each center ranged from 9 to 70. The averageaction and/or referrals received during their
calls were also evaluated. The companion ar- number of baseline assessments conducted
per counselor in each center ranged from 2ticle (Gould et al., this issue) addresses im-
mediate and intermediate outcomes for sui- to 33.
Baseline Cohort. Adult nonsuicidal in-cidal callers.
dividuals experiencing a crisis who called one
of the eight recruited TCS were the targeted
population for this study. Crises were definedMETHOD
as upset states precipitated by events with
which individuals currently felt unable toSample
cope. Between March 2003 and July 2004,
telephone crisis counselors conducted theTelephone Crisis Services. Eight TCS
were selected on the basis of organizational baseline assessment with 2,702 of the 5,168
eligible callers (52.3% participation rate). Ofstability (in operation at least 5 years), suffi-
cient call volume, quality assurance pro- these, 1,613 were crisis callers (25.9% male
Kalafat et al. 325
and 74.1% female). Gender was not coded at terviewers received passive or active refusals
at follow-up. Common reasons for counse-baseline for four crisis callers. The majority
of assessed crisis callers (90.1%; n = 1,357) lors not asking for consent for the follow-up
call were that the caller had to quickly termi-called the center’s local crisis hotline tele-
phone number and 149 (9.9%) called a na- nate the call, or hung up.
Crisis callers who participated in thetional 1-800 network that connected callers
to local crisis lines. follow-up assessment were similar to nonpar-
ticipants with regard to crisis state at the be-Of the 5,168 eligible callers, 2,466
(47.7%) callers were not assessed for the fol- ginning of the call and to changes in their
crisis state from the beginning to the end oflowing reasons: call volume too high (788),
callers’ suicide risk status too high (654), call- the call. However, crisis callers who were fol-
lowed were significantly more overwhelmeders refused/hung up (648), counselor thought
not appropriate to assess (226), and phone and received significantly more referrals from
counselors than crisis callers who were notproblems (150).
Out of the 87,459 calls received by the followed (Table 1).
participating counselors, 82,291 (94.1%) call-
ers met exclusion criteria and were not as- Measures
sessed. Of the excluded callers, 31,862
(38.7%) individuals called only for informa- An advisory board, consisting of ex-
perts in crisis lines, risk assessments, andtion and referral but were not in crisis;
16,664 (20.3%) were third-party callers also evaluation; and a Crisis Centers Directors
Board, consisting of the directors from sixnot in crisis; 13,986 (17.0%) were intoxicated
and/or belligerent callers; 12,619 (15.3%) telephone crisis services originally recruited
to participate, were created. With the assis-were frequent chronic callers; 2,732 (3.3%)
were minors; 2,381 (2.9%) were non-English tance of these two boards, and SAMHSA
staff, the client-centered outcomes werespeaking; and 2,167(2.6%) callers were not in
a mental state fit to complete the assessment. identified and final drafts of the baseline and
follow-up measures and the implementationFollow-Up Cohort. Between April
2003 and August 2004, of the 1,617 crisis plan were prepared for piloting. The final
baseline assessment also incorporated recom-callers who completed the baseline assess-
ment, follow-up assessments were conducted mendations received from counselors who
piloted the assessment with 103 callers inwith 801 (49.5%) callers (23.6% male and
76.4% female). Follow-up assessments were four centers. Manuals were developed to
train the crisis counselors and the follow-upconducted between 1 and 52 days from the
baseline assessment date, with the average interviewers.
Profile of Mood States: Modified (POMS-being 13.5 days. For these callers, the age
ranged from 18–85, and the mean was 37.6 M). Early work in the field of crisis inter-
vention included descriptions of the crisisyears. The ethnic distribution was 57.3%
White, 26.0% African American, 13.1% His- state based on observations of individuals’ re-
sponses to disasters, war, and other situa-panic, 1.4% other, 1.3% Native American,
and 1.0% Asian. Ethnicity was not coded for tional crises (Stein & Lambert, 1984). These
descriptions include cognitive componentsone caller.
For the 816 (50.5%) that did not par- such as confusion, overwhelmed, and con-
stricted problem solving; affective compo-ticipate in the follow-up, the reasons for not
participating were: 470 (57.6%) refused at nents such as anxiety, helplessness, and anger;
and cognitive-affective states such as hope-baseline, 124 (15.2%) gave the crisis counse-
lors invalid contact information, 69 (8.5%) lessness and depression. Halpern (1973) de-
veloped a self-report measure that consistedwere not asked if they wanted to receive a
follow-up call, and 153 (18.7%) gave consent of statements reflecting most of these states
and found that it reliably distinguished be-for follow-up contact but the follow-up in-
326 Nonsuicidal Crisis Caller Outcomes
TABLE 1
Crisis Callers: Comparison of Follow-up Participants and Nonparticipants
Followed Not Followed
(n = 801) (n = 816)
Time 1 Mean SD Mean SD
Total POMS-M 33.33 (11.12) 32.03 (11.77) F = 2.16 ns
Confusion 6.83 (3.30) 6.71 (3.34) F = 0.32 ns
Depression 7.74 (3.25) 7.62 (3.39) F = 0.23 ns
Anger 6.09 (3.56) 5.59 (3.62) F = 3.08 ns
Anxiety 7.52 (3.20) 7.35 (3.38) F = 0.96 ns
Helpless 2.42 (1.40) 2.30 (1.41) F = 3.44 ns
Overwhelmed 2.98 (1.19) 2.78 (1.32) F = 6.81**
Hopelessness 2.55 (1.03) 2.48 (1.04) F = 4.81 ns
Initiated Rescue Procedure 2 0.3% 1 0.1% χ
2
= .00 ns
Referral Given 541 68.0% 428 53.4% χ
2
= 34.73***
**p < .01; ***p < .001
tween individuals experiencing different types be administered, was also eliminated. This
reduced the number of items to 12. Two ad-of crises (e.g., interpersonal, divorce, be-
reavement) and a non-crisis comparison group. ditional items, helpless and overwhelmed, were
then added to the assessment to capture theFor the purposes of the present study, we
sought an instrument that described compo- words most commonly expressed by crisis
callers to describe how they are feeling. A to-nents of the crisis state and was brief, sensi-
tive to short-term changes, and psychometri- tal score on the POMS-M was the sum of
all 14 items. Callers were asked to rate theircally sound. The POMS-A met these criteria
and an adaptation of it was used to assess call- feelings on a 5-item scale (Not at all, A little,
Moderately, Quite a bit, Extremely) near the be-ers’ crisis state or level of distress. The
POMS has been utilized in hundreds of in- ginning of the call to the center, again at the
end of the call, and at the follow-up inter-vestigations to measure transient mood states
(McNair, Lorr, & Droppleman, 1992). The view.
Hopeles sn ess . Call er s were also assessedshortened version of the POMS-A (24 items)
is suitable for use with adolescents as well as for feelings of hopelessness by two sets of
questions. Callers were first asked, “To whatadults (Terry, Lane, Lane, & Keohane, 1999).
The POMS-A uses a “right now” time frame. degree do you feel that there is no hope for
improvement in your situation in the future?Based on factor analytic studies (McNair et
al., 1992; Norcross, Guadagnoli, & Pro- As you look into the future, do you see things
getting better in your life?” Responses werechaska, 1984; Rhoades, Grabowski, Elk, &
Cowan, 1993; Usala & Hertzog, 1989), six rated from 1 (Nothing will change, things will
stay bad)to5(Sure that the future will be bet-factors have been derived from the POMS
and POMS-A; namely, tension-anxiety, depres- ter). The response codes were recoded so that
higher scores indicated more hopelessness.sion-dejection, anger-hostility, fatigue-inertia,
confusion-bewilderment, and one positive state, The following question was, “To what extent
does the following belief, which I am aboutvigor-activity. To facilitate the use of this as-
sessment during a crisis call, the measure- to say, describe how you are feeling right
now? I don’t think I can go on.” Responsesment was shortened by excluding two factors,
fatigue-inertia and vigor-activity. The lowest were rated on the scale from not at all to ex-
tremely. Hopelessness was the average of theloading item, one for each of the domains to
Kalafat et al. 327
scores from both items. These questions Training Procedures
were asked at the beginning of the call to the
center and repeated at the end of the call and Center Staff. The research team trained
the crisis centers’ staff on the baseline admin-at the follow-up interview.
The remaining measures were asked istration protocols. In five centers, the re-
search team directly trained the counselors.only during the follow-up interview.
Plan of Action Compliance. This set of In the remaining three centers, the research
team trained one or more crisis center mem-questions assessed whether callers agreed
with and followed through with the plans of bers who then trained the centers’ counse-
lors. The suicide risk assessment protocol isaction developed by the crisis counselor with
the caller. Action plans ranged from “looking described in the companion paper (Gould et
al., this issue). For nonsuicide crisis calls,for a new job” to “taking a walk to calm
down.” If callers did not agree with the plan, counselors were instructed to conduct the
POMS and hopelessness assessments on allor if they did not completely follow through
with the plan, then their verbal responses as of the calls. Criteria for excluding calls from
the assessment were developed in collabora-to why they did not agree or follow through
were recorded as text responses. tion with center directors from the advisory
board. These were individuals who calledService Utilization and Compliance.
These questions assessed whether callers only for information and referral but were
not in crisis; third-party callers; intoxicatedagreed with and followed through with refer-
rals given to them. Callers were asked and/or belligerent callers; frequent chronic
callers; minors; non-English speaking callers;whether they remembered receiving refer-
rals, the type of referral(s) received (emer- and callers who were not in a mental state fit
to complete the assessment. Aside from in-gency services, mental health services, social
services, and information and referral ser- clusion of the POMS and hopelessness, we
did not promote any other changes in thevices), the extent of their agreement with the
referral(s), and the extent of follow through. centers’ usual crisis procedures or interven-
tions. Counselors were trained to ask ques-If callers did not agree with the referral, or if
they did not follow through with the referral, tions by incorporating them into their own
centers’ standard assessment and interventionthen their verbal responses as to why they did
not agree or follow through were recorded as procedures and helping styles. This assured
that the call would flow smoothly and nottext responses. These narrative responses
were recoded by two independent raters. Ex- feel like a structured interview. For example,
counselors were encouraged to use their ownamples of codes were, “services too far away,”
and “unable to pay for service.” language and style to ask questions, and to
use common crisis intervention language,Client Feedback on Call. Two open-
ended questions about what was helpful or such as “it sounds as if” or “I’m wondering.”
Also, counselors were trained to not asknot helpful about the call initiated the assess-
ment: “Thinking back to the call you placed questions but to just code the responses if a
caller spontaneously provided answers toto the crisis line, can you tell me how the call
was helpful to you?” “Can you tell me what questions. The training included role-playing
and discussions about what was or was notwas not helpful about the call?” Follow-up
interviewers made verbatim notes of callers’ working after each roleplay.
Follow-up Interviewers. To ensure in-responses to these questions. Twenty-one
close-ended questions followed the open- dependent follow-up assessments, these in-
terviewers were paid members of the projectended assessment and provided ratings in
three areas: helper interventions, emotion evaluation staff and not crisis center staff.
They had either telephone crisis counselingregulation, and overall efficacy, but the re-
sponses to the close-ended questions are the experience or equivalent clinical training and
experience. Training of follow-up interview-focus of another paper.
328 Nonsuicidal Crisis Caller Outcomes
ers included instructions on how to maintain between the counselor’s assessment and their
monitoring assessment and provided feed-client confidentiality during follow-up con-
tact; how to obtain informed consent; how to back to the counselor.
Intermediate Outcomes (Follow-up).administer the assessment in a compassionate
manner while retaining control of the inter- Eighteen follow-up interviewers participated
in the follow-up data collection. Interviewersview; and how to conference callers back to
the crisis center when they met criteria for were monitored during their first follow-up
assessment. The Project Coordinator moni-the required conference call as described in
the safety procedures section. Training was tored their audiotaped interviews and assess-
ments, and provided feedback to the inter-conducted through instruction and role-
playing. viewers in order to improve the quality of
their assessments.
Quality Control Procedures
Consent and Safety Procedures
Immediate Outcomes (Baseline). Sti-
pends were provided to each crisis center for Immediate Outcomes (Baseline). The
project was approved for a waiver of consenta staff member to function as a local data co-
ordinator under the supervision of the center for use of de-identified information gathered
during the baseline assessment. The assess-director. The local data coordinators re-
viewed completed assessments, provided on- ment procedures involved no more than min-
imal risk and were routine for telephone cri-going feedback and training to the counse-
lors, and sent de-identified baseline data and sis services. Compensation was not offered to
callers for completing the baseline assess-callers’ contact information to the Project
Coordinator. Local data coordinators were ment because it was included as part of the
intervention normally provided to callersalso responsible for reviewing the centers’
call records and comparing them to com- (such as a risk lethality assessment and asking
callers about their thoughts and feelings).pleted assessments to ensure that all eligible
callers were being assessed. If assessments Using an IRB-approved re-contact
consent script, counselors asked callers withwere not conducted with potentially eligible
callers, the coordinators reviewed the call re- whom they had conducted baseline assess-
ments if they wished to receive a follow-upcords for these callers with the crisis counse-
lors. The local data coordinators and the call from the research team in a week or two
to see if they were interested in participatingProject Coordinator communicated via tele-
phone and e-mail on a weekly basis to discuss in a follow-up assessment. To protect the
confident ial it y of calle rs during the re-contac tthe quality of the data collection and clarify
protocol procedures. efforts, counselors asked callers how and
when they wanted to be contacted, as well asFor additional quality control purposes
and to ensure the reliability of the baseline what type of message (if any) could be left
on an answering machine or with the personassessments, on-site silent monitoring was
conducted on approximately 10% of calls in picking up the telephone.
Intermedi ate Outcomes (Follow-up). Ac-each of the centers. Silent monitoring as-
sessed whether counselors were following the tive consent to participate in the follow-up
telephone survey, and for the research teamproject’s protocol. The monitors, drawn from
the centers’ local communities or from the to access callers’ baseline response and refer-
ral information, was obtained using an ap-centers’ staff, were hired by and reported to
the research team. As the monitor listened to proved telephone consent script at the start
of the follow-up call. A waiver of documenta-a call, s/he completed a baseline assessment
that was later compared to the assessment tion of consent was obtained so that consents
could be audiotaped rather than written. Forcompleted by the counselor. At the end of
the call, the monitor noted any discrepancies quality control purposes, approximately 10%
Kalafat et al. 329
of each follow-up interviewers’ assessments A repeated measures design was em-
ployed to examine changes over time, alwayswere audiotaped with callers’ consent to do
so. Callers participating in the follow-up as- employing center as the between subjects
factor. The measures were assessed at threesessment received a $25 money order.
Safety Procedures. The follow-up as- time points: near the beginning of the call
(Time 1), at the end of the call (Time 2), andsessment included criteria to determine
whether callers needed intervention, which at follow-up (Time 3). The repeated re-
sponse measures for the crisis callers were to-was defined as callers having made plans or
tried to hurt or kill themselves since speaking tal POMS-M score, the four component
scores derived from the POMS-M, andwith the center, or having serious intent to
die. The method for getting help to callers hopelessness.
Chi-square tests comparing males toconsisted of follow-up interviewers re-connect-
ing the caller to their crisis center while the females on problems mentioned at baseline
were performed. Crisis callers who were fol-interviewer remained on the line to ensure
that the caller was in communication with lowed were compared to those who were not
followed on baseline measures at the begin-the center. If callers were unable to partici-
pate in a call with their center immediately ning of the call, by means of univariate analy-
ses of variance. Interactions between follow-after completing their interviews, follow-up
interviewers obtained callers’ consent for the up status and changes from Time 1 to Time
2 were examined using two-way analyses ofcenter to contact them. In this last instance,
the follow-up interviewer contacted the cen- variance.
Immediate outcomes (Time 1 to Timeter and gave them the caller’s contact infor-
mation and details as to why the caller 2) and intermediate outcomes (Time 1 to
Time 2 to Time 3) were assessed by a re-needed intervention.
A confidentiality certificate was ob- peated measures design, including center as
the between subjects factor. Also assessed bytained from SAMHSA’s Depart men t of Health
and Human Service. The project’s protocol repeated measures were immediate and inter-
mediate outcomes by type of initial problem.was approved by Rutgers Graduate School of
Applied and Professional Psychology and the The statistical analyses were con-
ducted with SPSS statistical software (versionNew York State Psychiatric Institute/Colum-
bia University’s Institutional Review Boards. 12.0). Given the number of comparisons, re-
sults were considered significant at α<.001,
but results at α<.01 are presented in the ta-Analytic Strategy
bles.
The primar y samplin g unit of the study
was crisis center and the secondary sampling
RESULTS
unit was caller within center. Thus, we exam-
ined the extent to which within-center clus-
Presenting Problems
tering existed in order to determine whether
Callers presented at the center with a
this clustering variable needed to be included
variety of problems including abuse/violence
in the analyses. The sample clusters (center)
(10.8%), addictions (13.0%), base needs
had little impact on outcomes (distress [as
(18.7%), interpersonal problems (67.4%),
measured by the POMS-M], intent to die,
mental health (48.2%), physical health
and hopelessness) as indicated by the in-
(13.4%), work (9.9%), and other problems
traclass correlation coefficients, which were
(13.7%).
all close to zero (ranging from .004 to .05).
Therefore, the use of mixed-effects linear
Immediate Outcomes
models to account for the clustering variable
of center was unnecessary. Center was in- Callers’ distress, as assessed by the to-
tal score on the POMS-M, was significantlycluded as a covariate in the analyses.
330 Nonsuicidal Crisis Caller Outcomes
reduced from the beginning of the call (Time Suicidal Thoughts at Follow-up. At fol-
low-up, 94 (11.7%) of the 801 crisis callers1) to the end of the call (Time 2) (p < .001)
(see Table 2). There was also a significant re- reported that they had suicidal thoughts since
their call to the crisis center. When theseduction on the domains of the POMS-M as
well: confusion (p < .001), depression (p < callers were asked if they were having these
thoughts when they initially called the center,.001), anger (p < .001), anxiety (p < .001),
helpless (p < .001), and overwhelmed (p < 52 (55.3%) said yes, 35 (37.2%) said no, and
7 (7.4%) callers said they did not remember..001). There was also a significant reduction
in callers’ level of hopelessness (p < .001). Of the 52 callers who said they had suicidal
thoughts at baseline, 27 callers (51.9%) saidCallers with mental health problems were
significantly more depressed and hopeless they told the counselor about their thoughts,
17 (32.7%) said they did not tell the counse-than callers with nonmental health problems
(p < .001) (Table 3); however, there was no lor, and 8 (15.4%) said they did not remem-
ber. Compared to the crisis callers who didsignificant interaction between mental health
status and time. In other words, a mental not report any suicidal thoughts since their
initial call, the 94 crisis callers who reportedhealth problem did not modify the change
from Time 1 to Time 2. suicidal thoughts were significantly more dis-
tressed, as indicated by their total and com-
Intermediate Outcomes
ponent scores on the POMS-M and hope-
lessness scores at follow-up (Time 3) (Table
There was a significant reduction in
5). They were also significantly more de-
callers’ distress levels from the end of the call
pressed at the beginning of the baseline call
(Time 2) to follow-up (Time 3) as measured
(Time 1) and there was a tendency for these
by the total POMS-M score (p < .001) (see
callers to be more helpless and hopeless (p <
Table 4). Each of the individual scales of the
.01) at the baseline (Time 1) and more hope-
POMS-M and callers’ feelings of hopeless-
less (p < .01) at the end of calls (Time 2).
ness also showed significant reductions over
time (p < .001). Mental health problems did
Referrals
not modify the changes from the end of the
call (Time 2) to follow-up (Time 3), as indi- Out of the 1,617 callers who partici-
pated in the baseline assessment, 969 (59.9%)cated by the nonsignificant interaction effects
of problem by time. were given a new referral, of which 67.9%
TABLE 2
Crisis Callers: Immediate Outcomes (n = 1,617)
TIME 1 TIME 2
Main Effect
Mean SD Mean SD of Time
Total POMS-M 32.72 (11.46) 22.98 (12.54) F = 968.47***
Confusion 6.79 (3.31) 4.59 (3.08) F = 618.58***
Depression 7.67 (3.32) 5.54 (3.40) F = 582.55***
Anger 5.87 (3.60) 4.01 (3.36) F = 398.41***
Anxiety 7.44 (3.27) 5.32 (3.23) F = 471.23***
Helpless 2.36 (1.41) 1.55 (1.27) F = 352.74***
Overwhelmed 2.89 (1.26) 2.01 (1.29) F = 494.13***
Hopelessness 2.51 (1.04) 2.08 (0.90) F = 291.72***
Note. Significant differences over time for all measures at p <
.001.
***p < .001
Kalafat et al. 331
TABLE 3
Crisis Callers: Immediate Outcomes by Initial Problem
Mental Health Non-Mental Health
(n = 771) (n = 826)
Interaction
Effect of
Time 1 Time 2 Time 1 Time 2
Main Effect Problem
Mean SD Mean SD Mean SD Mean SD of Problem by Time
Total POMS-M 33.70 (11.25) 24.76 (12.79) 31.81 (11.56) 21.21 (12.03) F = 6.52** F = 0.03 ns
Confusion 6.79 (3.39) 4.88 (3.17) 6.80 (3.23) 4.33 (2.94) F = 2.55 ns F = 0.47 ns
Depression 8.18 (3.15) 6.14 (3.42) 7.19 (3.38) 4.97 (3.28) F = 26.12*** F = 2.22 ns
Anger 5.80 (3.59) 4.12 (3.47) 5.90 (3.60) 3.89 (3.26) F = 3.28 ns F = 1.73 ns
Anxiety 7.72 (3.18) 5.86 (3.26) 7.22 (3.33) 4.85 (3.13) F = 9.61** F = 0.07 ns
Helpless 2.44 (1.38) 1.63 (1.26) 2.30 (1.43) 1.48 (1.27) F = 5.07 ns F = 4.56 ns
Overwhelmed 2.97 (1.21) 2.17 (1.30) 2.82 (1.29) 1.87 (1.26) F = 6.40 ns F = 0.23 ns
Hopelessness 2.62 (1.06) 2.20 (0.90) 2.41 (1.01) 1.97 (0.89) F = 16.27*** F = 7.07**
**p < .01; ***p < .001
332 Nonsuicidal Crisis Caller Outcomes
TABLE 4
Crisis Callers: Intermediate Outcomes (n = 801)
Time 1 Time 2 Time 3
Main Effect T2–T3
Mean SD Mean SD Mean SD of Time Contrast
Total POMS-M 33.37 (11.07) 24.01 (12.67) 17.30 (13.84) F = 258.90*** F = 61.56***
Confusion 6.85 (3.29) 4.76 (3.17) 3.50 (3.23) F = 177.13*** F = 36.60***
Depression 7.69 (3.25) 5.71 (3.46) 3.76 (3.48) F = 222.03*** F = 76.43***
Anger 6.11 (3.56) 4.22 (3.51) 3.02 (3.27) F = 122.69*** F = 25.33***
Anxiety 7.50 (3.19) 5.60 (3.27) 4.45 (3.46) F = 133.12*** F = 25.36***
Helpless 2.43 (1.40) 1.62 (1.30) 1.02 (1.31) F = 151.42*** F = 56.98***
Overwhelmed 2.99 (1.19) 2.13 (1.30) 1.55 (1.49) F = 198.56*** F = 46.33***
Hopelessness 2.55 (1.03) 2.10 (0.88) 1.75 (0.85) F = 214.92*** F = 64.27***
Note. Significant differences over time for all measures at p < .001.
***p < .001
(658) were to mental health resources. An ad- included such strategies as discussing the
problem with a partner, calling a lawyer, orditional 135 (8.3%) callers were referred back
to their current therapist or services. Of the engaging in a relaxing and/or diversionary
activity. Out of the 801 crisis callers who par-801 callers who participated in the follow-up,
541 (67.5%) were given a new referral at ticipated in the follow-up, counselors devel-
oped plans with 464 (57.9%). At follow-up,baseline, of which 72.5% (392) were to men-
tal health resources. An additional 75 (9.4%) 369 (79.5%) callers recalled the plan and of
these, 160 (43.4%) callers completed “all” ofcallers were referred back to their current
therapist or services. The overall referral rate the plan, 47 (12.7%) callers completed
“most”, 54 (14.6%) completed “some,” 72for those who participated in the baseline was
68.3% and the rate of referral for those who (19.5%) said the plan was still “in process,”
and 35 (9.5%) callers had not carried out anyparticipated in the follow-up was 76.9%.
Callers’ type of problem had little im- of the plan. The extent of follow through was
not coded for one caller (.3%).pact on whether or not a new referral was
given. The referral rate by callers’ type of
problem ranged from 53.1% to 69.6%; how- Re-Contact with the Center
ever, callers with mental health problems and
or base needs problems had the highest refer- Of the 801 crisis callers who partici-
pated in the follow-up, 186 (23.2%) callersral rate of 69.6%. Of the 392 follow-up crisis
callers who were given a new mental health had re-contact with the crisis center after
their initial call with the center. Because thereferral, 33.2% had kept or made an appoint-
ment with a mental health service in the pe- centers categorized these as nonsuicide crisis
calls, it is most likely that the callers ratherriod between the initial call to the center and
the follow-up assessment. than the centers initiated the re-contact. Of
those who had re-contact with the center, 96
(51.6%) callers had one additional contact,Plans of Action
32 (17.2%) had two contacts, 27 (14.5%) had
three contacts, 6 (3.2%) had four contacts, 17During the calls, crisis counselors de-
veloped “plans of action” with callers. These (9.1%) had between 5 and 15 contacts, and 8
(4.3%) callers did not remember the numberwere plans that callers agreed to follow
through with to try to help themselves. Plans of times. More callers who had re-contact
Kalafat et al. 333
TABLE 5
Crisis Callers Endorsing Suicidal Thoughts at Follow Up, Compared to Crisis Callers Not Endorsing
Suicidal Thoughts at Follow Up
Endorsing Suicidal Not Endorsing Comparison of
Thoughts Suicidal Thoughts Means at
(n = 94) (n = 706) Follow-Up
T1 T2 T3 T1 T2 T3 T1 T2 T3
Total POMS-M
Mean 35.93 25.85 25.41 32.98 23.81 16.16 2.36 ns 1.38 ns 6.22***
(SD) (9.56) (11.33) (15.20) (11.29) (12.89) (13.24)
Confusion
Mean 7.51 5.07 5.17 6.74 4.69 3.21 2.11 ns 1.05 ns 5.63***
(SD) (2.95) (2.73) (3.72) (3.34) (3.22) (3.08)
Depression
Mean 8.88 6.45 5.88 7.58 5.63 3.47 3.66*** 2.14 ns 6.45***
(SD) (2.55) (2.94) (3.76) (3.30) (3.50) (3.34)
Anger
Mean 6.13 4.03 4.52 6.10 4.25 2.83 .09 ns −.57 ns 4.77***
(SD) (3.51) (3.28) (3.73) (3.56) (3.55) (3.14)
Anxiety
Mean 7.98 6.09 5.76 7.45 5.54 4.25 1.50 ns 1.50 ns 4.00***
(SD) (2.97) (3.14) (3.65) (3.22) (3.28) (3.40)
Helpless
Mean 2.77 1.72 1.73 2.37 1.62 .93 2.59** .67 ns 5.65***
(SD) (1.18) (1.23) (1.44) (1.43) (1.31) (1.27)
Overwhelmed
Mean 3.06 2.09 2.34 2.97 2.14 1.45 .75 ns −.34 ns 5.52***
(SD) (1.08) (1.20) (1.54) (1.21) (1.31) (1.46)
Hopelessness
Mean 2.85 2.34 2.45 2.51 2.07 1.66 3.01** 2.73** 8.74***
(SD) (1.03) (0.93) (0.98) (1.03) (0.88) (0.79)
**p < .01; ***p < .001
with the center had thoughts about killing Caller Feedback
themselves since their initial call to the cen-
ter (18.8%) than those who had not re- At follow-up, 801 crisis callers pro-
vided a total of 1,345 responses to the posi-contacted the center (9.6%; χ
2
= 10.69, p <
.001). On the POMS at the follow-up call tive question and 145 responses to the nega-
tive question. Fifteen crisis callers said(T
3
), those who had re-contact with the cen-
ter as compared to those who had not were nothing was helpful about the call.
Two raters, both of whom were experi-significantly more hopeless (t = 3.48, p <
.001), and there were trends toward their be- enced crisis counselors, independently coded
the positive responses into 17 a priori catego-ing more anxious (t = 2.90, p < .004), de-
pressed (t = 3.01, p < .003), confused (t = ries, plus six additional categories that were
developed to account for the responses gen-2.97, p < .003), and overall more distressed as
measured by the total POMS score (t = 3.05, erated through content analysis. Negative
statements were coded into 15 categoriesp < .002).
334 Nonsuicidal Crisis Caller Outcomes
generated from content analysis of the re- the hypothesized role of telephone crisis in-
tervention in attenuating crisis states. Whilesponses. The raters agreed on 86% of their
coding of positive statements and 74% of demand characteristics may play a role in
callers’ reports to helpers at the end of calls,their coding of negative statements. Dis-
agreements, most of which consisted of one the continued or additional reductions in dis-
tress reported by callers to independent eval-of the raters failing to rate a statement, were
resolved through discussion. uators at follow-up attenuates this concern to
a degree. This finding also suggests thatMany categories accounted for a small
percent of the responses. The top six catego- changes occurring during the calls are not
transitory phenomena and may set the stageries of positive responses (>6% of responses)
described empathic helpers (7.8% of re- for improved coping with crises.
Nearly a quarter of callers had re-sponses; 13.1% of callers) who listened and al-
lowed the callers to talk about their concerns contact with the centers for one or more
calls. Multiple contacts by callers to crisis(23.4%; 39.2%), helped them to calm down
(9.2%; 15.4%), think more clearly (6.8%; centers are a complex phenomenon. For
some callers, these may represent a “safety11.4%), and provided options for dealing with
their concerns (15.4%; 25.8%). The services net” when local communities lack sufficient
support resources, or simply indicate that cri-were described as readily available with help-
ers willing to stay on the line as long as sis services may require more than one con-
tact to meet the needs of callers. For otherneeded (available, patient , 6.7%; 11.2%).
The most common problem noted by callers, multiple calls to centers may indicate
inappropriate reliance on the service in placecallers concerned the referrals provided by
crisis staff (23.2% of responses; 5.6% of call- of follow through with other options. Fur-
ther research is necessary to clarify this phe-ers). Some of the referrals were not appro-
priate for the caller’s problem, but most of nomenon, as repeat callers comprise a sub-
stantial proportion of calls to some centers.the difficulties with referrals were due to the
agencies to which callers were referred. The Previous research has provided some
support for the helping model espoused bynature of these difficulties included cost,
waiting lists, and unhelpful responses. The many telephone crisis services (Echterling &
Hartsough, 1989; Echterling et al., 1980;next most frequent concerns were about in-
adequate solutions to problems. The callers Young, 1989), and the caller feedback ob-
tained in this study also conforms to the pro-raised concerns that they weren’t given any
help on how to solve their problem: “they posed elements of telephone intervention.
This feedback described available, empathicjust comforted me” (10.8%; 2.6%); or, were
given unhelpful suggestions/solutions (10.3%; helpers who listened to callers and let them
talk, helped them to calm down, and pro-2.5%) (e.g., “He said things I already know”).
Callers also indicated that crisis staff asked vided options for addressing their concerns.
Reduction in distress is one probabletoo many questions (10.8%; 2.6%); yet the
“calls were too short” (8.3%; 2.0%). Callers mediator of positive resolution of crises. An-
other element of crisis intervention is the de-also identified unhelpful characteristics of
some counselors (e.g., condescending, not velopment of a plan of action for addressing
the caller’s concern and/or the provision of aconcerned, abrupt) (8.8%; 2.1%).
referral to an agency that can further assist
the caller. Of the 801 followed callers, 392
(48.9%) were given mental health referrals atDISCUSSION
baseline, which is the largest category of re-
ferrals given. Of these, 33.2% had made orSignificant changes from the begin-
ning to the end of the calls in desirable direc- kept an appointment since their original call.
This is lower than the modal 50% followtions in affective and cognitive variables that
characterize crisis states are consistent with through rate reported for earlier studies
Kalafat et al. 335
(Stein & Lambert, 1984), although the types Finally, there is concern about the
finding at follow-up that 94 (11.7% of 801of referrals were not specified in these stud-
ies. Reasons for lack of follow through were crisis callers followed up) reported that they
had suicidal thoughts since their original callsnot obtained in the current study, although
some information is available from 45 callers to the centers, that 52 of these individuals
were having these thoughts when they calledwho, in response to the open-ended ques-
tions as to what was not helpful about the the center, and 17 said they had told the
counselor about these thoughts. This has twocall, reported problems with the referrals
provided by crisis services. These included implications for crisis services: first, policies
and procedures must be developed regardingreferrals that were not appropriate for the
callers’ concerns and problems with the assessment of suicidal risk for all crisis callers
or for a clearly specified type of crisis caller;agency to which the caller was referred, such
as cost and long waiting times. At least one and, second, training must be developed that
enhances the ability and inclination to sys-previous study found that referral follow
through can be improved by provision of in- tematically assess for suicide risk. The need
to systematically conduct evidence-based riskcentives for staff for more conscientious re-
ferrals (Paul & Turner, 1976). Thus, training assessments is attested to by the finding that
the suicidality of a substantial number of cri-may ameliorate this problem to some degree.
However, as crisis services expand beyond sis callers was missed during the crisis inter-
vention and that they remained suicidal attheir own communities to statewide or multi-
state service areas, they need to maintain follow-up. It should be noted that failure to
conduct risk assessments or pursue patients’more extensive current and accurate referral
databases as well. It may also be possible that suicidal communications has also been found
among professional mental health providersthe attenuation of crisis states achieved by
crisis workers may have reduced the distress- (Bongar, Maris, Berman, & Litman, 1998;
Coombs et al., 1992) and primary care physi-mediated motivation to seek further help.
Further research is necessary to test this hy- cians (Adamek & Kaplan, 2000; Williams et
al., 1999). These findings for mental healthpothesis of unintended side effects of crisis
intervention. and medical providers have been attributed
to the documented lack of training in suicideCrisis counselors developed plans of
action with 464 (57.9%) callers for address- risk assessment and management for these
professionals (Bongar, 2002; Williams et al.,ing their concerns. Of these, 20.5% did not
recall their plans, 43.4% reporting having 1999). This reinforces the need for training
in this area.completed the entire plan, 46.8% reported
various stages of progress on their plan, and The present study has several advan-
tages for examining the effectiveness of tele-9.4% had not carried out any of the plan.
This is a more encouraging follow through phone crisis services. First, this study pro-
vided an empirical evaluation of crisis hotlinerate than the referral rate. There are several
possible reasons for this, including the possi- services using a broader range of potential
beneficial effects than used in an earlier gen-bility that the plans are more subject to con-
trol by the callers and/or may require less eration of evaluation studies. Conceptually
and/or empirically based variables associatedtravel and cost. Also, plans of action may be
more likely to be developed through a collab- with crisis states of callers were selected as
immediate and intermediate outcomes. Sec-orative process. A study by Echterling and
Hartsough (1989) provides some support for ond, while the inclusion of a control condi-
tion, for example, a placebo call protocol, wasthis. They found a positive relationship be-
tween problem solving involving the develop- unethical, the repeated measures design al-
lowed each caller to act as his/her own con-ment of action plans that occurred near the
end of calls, as opposed to earlier phases of the trol in the assessment of the client-centered
outcomes. Third, the follow-up assessmentcalls, and positive call outcomes at follow-up.
336 Nonsuicidal Crisis Caller Outcomes
provided invaluable information on referral coordinators, who were reimbursed by the
project grant to ensure appropriate data col-follow through and allowed for an evaluation
of the callers by interviewers independent of lection, may also represent an additional re-
source that may not be available in manythe crisis centers. Fourth, in contrast with
most earlier studies, a multisite methodology centers. Third, while selection biases may
also exist with regard to the callers who werewas employed, which may increase the gen-
eralizability of the findings. While not a rep- followed, for crisis callers the concern about
possible positive selection bias among callersresentative sample of U.S. crisis centers, a
geographically diverse set of centers with who consented to follow-up is attenuated by
the finding that there were almost no differ-varied counselor characteristics (e.g., volun-
teer or paid; lay or professional) was em- ences between the baseline sample that was
not followed up and the follow-up sample inployed and yielded the largest sample of call-
ers in nonsuicidal or suicidal crises studies to levels of distress at the beginning of their
calls nor in changes from the beginning todate.
The study has important limitations. end of the calls. The lone exception was that
followed callers were significantly more over-First, the study was uncontrolled, as men-
tioned above, because of ethical concerns whelmed at the start of their calls than non-
followed callers. Fourth, telephone crisis ser-about compromising the clinical services pro-
vided to persons in crisis. Second, selection vices did not routinely collect demographic
information, such as age and ethnicity, onbiases exist with regard to the centers and
counselors who participated. The participat- callers, which precluded our ability to exam-
ine the specificity of the findings for differenting centers and counselors had to be amena-
ble to implementing a series of questions populations of users.
In sum, the information yielded by thisabout the caller’s current emotional state,
which was not compatible with some centers’ study is consistent with the effectiveness of
telephone crisis services in reducing the crisisor counselors’ helping model. Furthermore,
the implementation of the research protocol state of callers. However, the lack of a control
condition makes it difficult to definitively at-may have influenced the nature of the inter-
action between the helper and the caller. An- tribute the improvements in crisis state to the
crisis intervention. Based on the feedbackecdotal reports from crisis staff were mixed
in that some found the questions to be some- provided by callers, as well as the record of
referrals and action plans, these centers ap-what intrusive, while others indicated that it
facilitated their assessment of the caller’s pear to be providing callers with opportuni-
ties to problem solve and identify resourcesstate, helped the callers to clarify their feel-
ings, and helped the callers and crisis workers for addressing their concerns. The results
also highlight the need to systematically em-to see the progress achieved during the call.
In any case, the results can only be general- ploy reliable and valid risk assessments on
crisis calls and to enhance successful referralized to an intervention model that incorpo-
rates some direct assessments of callers’ men- of callers in need to appropriate services.
tal state. The efforts of the local data
REFERENCES
Adamek, M. E., & Kaplan, M. S. (2000).
Baldwin, B. A. (1979). Crisis intervention:
Caring for depressed and suicidal older patients:
An overview of theory and practice. Counseling
A survey of physicians and nurse practitioners. In-
Psychologist, 8, 43–52.
ternational Journal of Psychiatric Medicine, 30, 111–
Bleach, G., & Claiborn, W. L. (1974).
125.
Initial evaluation of hotline telephone crisis cen-
Apsler, R., & Hodas, M. (1975). Evaluat-
ters. Community Mental Health Journal, 4, 387–
ing hotlines with simulated calls. Crisis Interven-
tion, 6, 14–21. 394.
Kalafat et al. 337
Bongar, B. (2002). The suicidal patient: San Diego, CA: Educational and Industrial Test-
ing Service.
Clinical and legal standards of care. Washington,
Norcross, J. C., Guadagnoli, E., &
DC: American Psychological Association.
Prochaska, J. O. (1984). Factor structure of the
Bongar, B., Maris, R. W., Berman, A. L.,
Profile of Mood States (POMS): Two partial repli-
& Litman, R. E. (1998). Outpatient standards of
cations. Journal of Clinical Psychology, 40, 1270–
care and the suicidal patient. In B. Bongar, A. Ber-
1277.
man, R. Maris, M. Silverman, E. Harris, & W.
O’Donnell, J. M., & George, K. (1977).
Packman (Eds.), Risk management with suicidal pa-
The use of volunteers in a community mental
tients (pp. 4–33). New York: Guilford.
health center emergency and reception service: A
Brockopp, G. W. (1973). Crisis interven-
comparative study of professional and lay tele-
tion: Theory, process, and practice. In D. Lester
phone counseling. Community Mental Health Jour-
& G. W. Brockopp (Eds.), Crisis intervention and
nal, 1, 3–12.
counseling by telephone (pp. 89–104). Springfield,
Paul, T. W., & Turner, A. J. (1976).
IL: Charles C. Thomas.
Evaluating the crisis service of a community men-
Buchta, R., Wetzel, R. D., Reich, T.,
tal health center. American Journal of Community
Butler, F., & Fuller, D. (1973). The effect of
Psychology, 4, 303–308.
direct contact with referred crisis center clients on
Rapoport, L. (1965). The state of crisis:
outcome success rates. Journal of Community Psy-
Some theoretical considerations. In H. J. Parad
chology, 1, 395–396.
(Ed.), Crisis intervention: Selected readings (pp. 2–
Caplan, G. (1964). Principles of preventive
31). New York: Family Service Association of
psychiatry. New York: Basic Books.
America.
Caruthers, J. E., & Inslee, L. J. (1974).
Rhoades, H. M., Grabowski, J., Elk, R.,
Level of empathic understanding offered by vol-
& Cowan, K. (1993). Factor stationarity and in-
unteer telephone services. Journal of Counseling
variance of the POMS in cocaine patients. Psycho-
Psychology, 21, 274–276.
pharmacolgical Bulletin, 29, 263–267.
Coombs, D. W., Miller, H. L., Alarcon,
Rogers , C. R. (1957). The nec ess ar y and
R. D., Herlihy, C., Lee, J. M., & Morrison,
sufficient condi tio ns of therapeutic personalit y
D. P. (1992). Presuicide attempt communications
change . Journal of Consulti ng Psych ol ogy, 21, 95–103.
between parasuicides and consulted caregivers.
Slaikeu, K. A., Tulkin, S. R., & Speer,
Suicide and Life-Threatening Behavior, 22, 289–302.
D. C. (1975). Process and outcome in the evalua-
Echterling, L. G., & Hartsough, D. M.
tion of telephone counseling referrals. Journal of
(1989). Phases of helping in successful crisis tele-
Consulting and Clinical Psychology, 43, 700–707.
phone calls. Journal of Community Psychology, 17,
Slaikeu, K. A., & Willis, M. A. (1978).
249–257.
Caller feedback on counselor performance in tele-
Echterling, L. G., Hartsough, D. M., &
phone crisis intervention: A follow-up study. Crisis
Zarle, T. H. (1980). Testing a model for the pro-
Intervention, 9, 42–49.
cess of telephone crisis intervention. American
Stein, D. M., & Lambert, M. J. (1984).
Journal of Community Psychology, 8, 715–725.
Telephone counseling and crisis intervention: A
Farberow, N. L., Heilig, S. M., & Lit-
review. American Journal of Community Psychology,
man, R. E. (1968). Techniques in crisis intervention:
12, 101–126.
A training manual. Los Angeles: Suicide Preven-
Terry, P. C., Lane, A. M., Lane, H. J., &
tion Center.
Keohane, L. (1999). Development and validation
France, K. (1975). Evaluation of lay volun-
of a mood measure for adolescents. Journal of
teer crisis telephone workers. American Journal of
Sports Science, 17, 861–872.
Community Psychology, 3, 197–219.
Usala, P. D., & Hertzog, C. (1989). Mea-
Gould, M. S., Kalafat, J., Harris
surement of affective states in adults: Evaluation
Munfakh, J. L., & Kleinman, M. (this issue). An
of an adjective rating scale instrument. Research on
evaluation of crisis hotline outcomes. Part II: Sui-
Aging, 11, 403–426.
cidal callers. Suicide and Life-Threatening Behavior,
Williams, J. W., Rost, K., Dietrich,
37, 000–000.
A. J., Ciotti, M. C., Zyzanski, S. J., & Cornell,
Halpern, H. A. (1973). Crisis theory: A
J. (1999). Primary care physicians’ approach to de-
definitial study. Community Mental Health Journal,
pressive disorders: Effects of physician specialty
9, 342–349.
and practice structure. Archives of Family Medicine,
Kalafat, J., Boroto, D. R., & France, K.
8, 58–67.
(1979). Relationships among experience level and
Young, R. (1989). Helpful behaviors in the
value orientation and the performance of parapro-
crisis center call. Journal of Community Psychology,
fessional telephone counselors. American Journal
17, 70–77.
of Community Psychology, 7, 167–180.
McNair, D., Lorr, M., & Droppleman, Manuscript Received: October 1, 2006
Revision Accepted: January 23, 2007L. F. (1992). POMS manual: Profile of Mood States.
