Rapid Essay Researchers We offer fast essays with the best content
read the article that’s below and answer the 9 questions that’s attatched
Systematic Study of Structured Diagnostic
Procedures in Outpatient Psychiatric
Rehabilitation:
A Three-year, Three-cohort Study of the
Stability of Psychiatric Diagnoses
by RAYMOND KOTWICKI, MD, AND PHILIP D. HARVEY, PhD
Dr. Kotwicki is with Skyland Trail and Emory University in Atlanta, Georgia; and Dr. Harvey is with the University of Miami Miller School of
Medicine, Miami, Florida.
Innov Clin Neurosci. 2013;10(5–6):14–19
FUNDING: This research was funded by Skyland Trail.
FINANCIAL DISCLOSURES: Dr. Kotwicki is the Medical Director of Skyland Trail. He reports no other conflicts of interest. Dr. Harvey is a member of
the National Advisory board of Skyland Trail and is compensated for this service.
ADDRESS CORRESPONDENCE TO: Raymond Kotwicki, MD; E-mail: rkotwicki@skylandtrail.org
KEY WORDS: Bipolar disorder, psychosis, structured diagnoses, validity
ABSTRACT clinician diagnoses and those appears to be a reasonable trade-off
Background. Psychiatric generated by structured interviews. between brevity and accuracy
diagnoses are important for The same three interviewers through the use of the MINI
treatment planning. There are a examined all patients in all three compared to the SCID, with
number of current challenges in the phases of the study. substantial improvements in stability
area of psychiatric diagnosis with Results. Admission and discharge of diagnoses compared to clinician
important treatment implications. In diagnoses were available for 313 diagnoses. Clinical diagnoses were
this study, we examined the cases. Diagnoses generated with the minimally overlapping with the
differential usefulness of two semi- unstructured procedure were results of structured diagnoses,
structured interviews of differing changed by discharge 74 percent of suggesting that structured
length compared to clinical diagnoses the time, compared to four percent assessment, particularly early in the
for generation of diagnoses that did for SCID diagnoses and 11 percent illness or in short term treatment
not require modification over the for MINI diagnoses. Referring settings, may improve treatment
course of treatment. clinician diagnoses were disconfirmed planning.
Methods. We performed a three- in Years 2 and 3 in 56 percent of
year, three-cohort study at an SCID cases and 44 percent of MINI INTRODUCTION
outpatient psychiatric rehabilitation cases. The distinctions between The reliability of psychiatric
facility, comparing the stability of unipolar and bipolar disorders were diagnoses has improved markedly
admission diagnoses when generated particular points of disagreement, since the introduction of structured
by unstructured procedures relying with similar rates of under and over- psychiatric interviews.1 These
on referring clinician diagnosis, the diagnosis of bipolar disorder. The rate interviews were first developed in the
SCID, and the MINI. We examined of confirmation of referring clinician late 1960s2 and were fine tuned3 up
changes in diagnoses from admission diagnoses of schizoaffective disorder through the time of the introduction
to discharge (averaging 13 weeks) was 10 percent with the SCID and 11 of the the Diagnostic and Statistical
and, during the second two years, percent with the MINI. Manual of Mental Disorders, Third
convergence between referring Discussion. In this setting, there Edition (DSM-III)4 in 1980. At the
14 Innovations in CLINICAL NEUROSCIENCE [VOLUME 10, NUMBER 5–6, MAY–JUNE 2013]
same time, the use of these multiple, newly indicated treatments patients accurate diagnosis is
structured interviews is still not and associated advertising. In important. This is particularly
common in everyday clinical practice, addition, an increased appreciation of relevant to time-limited treatment. As
with most use in research settings. It the fact that bipolar disorders can be interventions such as day treatment
is not clear how much the application marked by brief episodes of or other rehabilitation therapies may
of such interviews would impact the hypomania rather than full manic be approved by insurance payers for
reliability and validity of diagnoses in episodes has increased the challenge delivery only for finite periods,
clinical practice settings, but it seems in discrimination between bipolar and inaccurate targeting of treatment
likely that there are certain unipolar mood disorders. We know interventions early on could lead to
circumstances where the increase in that distinguishing unipolar therapeutic interventions being
validity would be quite substantial. depression and bipolar illness has applied for relatively abbreviated and
The importance of collection of valid socioeconomic and functional potentially inefficacious periods.
assessment data through structured implications.12 Correspondingly, Thus, early identification of the
assessment procedures is contemporary diagnostic trends may eventual diagnosis can lead to
compounded by the problems in self- also incorrectly shape referring enhanced ability to deliver
report seen in multiple psychiatric diagnoses when patients initially appropriate treatments for a larger
conditions;5–7 questionnaire or present for treatment. For instance, proportion of the time allowed. In
checklist methods that do not contain in previous years the concept of this context, stability of diagnoses
interaction and observation with an schizophrenia was expanded to over time reflects an important
interviewer are clearly subject to include a variety of conditions component of the validity of these
these concerns. outside the current boundaries, such diagnoses while it is admittedly not
While we have recently shown in a as “pseudo-neurotic schizophrenia;”13 the only important aspect.
literature review8 that established there is a controversy about whether This paper presents the results of
schizophrenia can be diagnosed by current concepts of mood spectrum a systematic study of the usefulness
clinicians with high degrees of conditions are overly broad as well. of structured psychiatric interviews.
concordance with the results of There are several benefits of In a three-year, three-cohort,
structured psychiatric interviews, systematic collection of diagnostic consecutive-admission study, we
there are still multiple diagnostic data in everyday practice. There are examined psychiatric diagnoses that
challenges. Patients with multiple, suggestions that certain conditions, were generated through unstructured
early-course conditions, even such as bipolar disorder, are both clinical interviews and reliance on
schizophrenia, often have diagnoses over-diagnosed14,15 and frequently referral source diagnoses (Year 1),
that change even when initially missed16,17 in clinical settings. The and two different psychiatric
generated with structured most frequent suggestion to remedy interviews that varied in their length
procedures.9,10 Psychiatric interviews this situation is a structured of administration (Years 2 and 3).
vary in their focus (Axis-I vs. Axis-II), psychiatric interview. In fact, in the This study was performed at an
in their length, and in their Pogge et al14 and Zimmerman et al15 outpatient psychiatric rehabilitation
assessment of the patient alone studies, using a structured interview center that largely focuses on early
versus symptoms in their relatives. revealed over-diagnosis of bipolar course patients (mean age=24) and
Structured interviews can require disorder in adolescent and adults included three years of consecutive
substantial time commitments and found to have major depression. admissions from similar referral
can require considerable training in Presumptive diagnoses of sources, where the assessment
order to be accurately employed. posttraumatic stress disorder (PTSD) procedure was systematically
Secular trends and patient are often generated on the basis of changed at one-year periods with the
expectations may also impact trauma exposure, without a same admission staff in place across
presumed diagnoses when new systematic assessment of the other the three years. We used the
patients present for treatment in required symptoms.18 Schizoaffective Structured Clinical Interview for the
community mental health settings. disorder is commonly diagnosed in DSM (SCID)22 for the second year of
Some of this variation may be due to clinical practice,19 but the diagnosis the study and the MINI International
exposure of potential patients to has been argued to lack reliability20 Neuropsychiatric Interview (MINI)23
media or internet information, which and intrinsic clinical validity.21 for the third. Stability of diagnoses
may shape their opinions of their Managed care companies are was indexed through the number of
diagnoses. Bipolar disorder, for often interested in matching changes in diagnosis suggested by the
instance, has seen a marked increase treatments to diagnoses and may clinical staff during the course of the
in terms of its diagnosed prevalence refuse to reimburse for treatments patient’s treatment based on real-
in the last 20 years, after 40 years of that are not approved for specific time observations and the results of
stability in diagnostic prevalence,11 indications, suggesting that in order the therapeutic process. For cases in
with this increase corresponding with to offer suitable treatments to Years 2 and 3, we also compared the
TABLE 1. Diagnoses at admission over a three-year, three-cohort prospective study as they were primary. Demographic
data, including admission diagnoses,
DIAGNOSTIC METHOD (YEAR)
are presented in Table 1. As can be
CLINICAL (ONE) SCID (TWO) MINI (THREE) seen in the table, the ages of the
N cases declined slightly each year and
110 101 110 there was a slight shift in the
M SD M SD M SD diagnostic distribution.
Procedure. The same three
Age 36.4 12.6 34.1 13.4 31.9 11.8 admission staff members participated
Length of stay (days) 90.6 66.6 86.7 75.2 81.8 51.2 in all three years of this study, which
started October 1, 2008. In year one,
Gender (% male) 62 59 50 all referrals for admission to the
Axis-I admission diagnosis
treatment center received a clinical
diagnosis based on an interview at
Bipolar 46 38 50 admission and information provided
by the referral source. Throughout
Major depression 30 38 40
that year, the presumed “working
Schizoaffective 11 1 1 diagnosis” was the referral diagnosis
accompanied by an unstructured
Schizophrenia 6 16 7
interview that occurred within 48
Anxiety (includes PTSD) 6 4 10 hours of the patients’ admission. In a
pre-planned study, the three staff
Other 1 4 2
members were trained by an
experienced psychiatric
referral source diagnosis for the informed consent for the analyses diagnostician. During Year 2, these
patients to the diagnosis generated performed in this study. Patients with same staff members, after training,
with a structured psychiatric a primary diagnosis of a substance interviewed all candidates for
interview. Our hypothesis was that use disorder or personality disorders admission with the SCID. Interview
both of the structured interviews were excluded from admission due to training consisted of observed
would be superior for generating regulatory issues during this time interviews, joint ratings, and
stable diagnoses to both clinical period. Dual-diagnoses patients as consensus discussion of a series of
judgments and referral diagnoses well as patients who had concomitant cases not included in these analyses.
based on unstructured clinical (but not primary) personality After one year of use of the SCID, a
observation. We were particularly disorders were included in analyses. third year of admissions were all
interested in whether the The same three experienced, interviewed and diagnosed with the
considerably more abbreviated MINI master-level, admission staff MINI interview using identical
would yield the same diagnostic members participated over all three training procedures.
stability, compared to the lengthier years. At the beginning of the study, There are several other important
SCID, in these patients. these staff members had a minimum features of this design. In order for
of three years of experience and an the procedure to simulate the reality
METHODS average of 5. Cases were distributed of clinical practice, we did not
Participants. Research sequentially across the three raters perform extensive assessment of
participants consisted of three years after referral to the treatment facility. inter-rater reliability after the
of consecutive admissions to a These staff members were not initiation of the project. Instead, our
private, nonprofit, psychiatric involved in the treatment of the goal was to determine if using a semi-
rehabilitation facility. All admissions patients and did not have input into structured interview and a trained
were examined; cases who were any subsequent treatment decisions. rater would generate stable
screened for admission but who did Further, the clinical staff members diagnoses. To examine this question,
not receive services were not treating the patients were not we compared the rate of clinician
analyzed. All data were archived in a informed of the plans to evaluate change of rater-generated admission
database and examined anonymously. diagnostic stability as an outcome diagnoses across the three raters.
Patients signed a general consent measure in the study. The reporting Thus, the outcome was diagnostic
form for their data to be examined of the diagnoses consisted of the axis stability across raters within rating
anonymously and the Emory I and axis II diagnostic impressions method and not agreement on a
University Internal Review Board which were entered into the specific diagnosis such as bipolar
approved this study with expedited electronic medical records. For this disorder or schizophrenia. Patients
review and did not require signed study, we focused on axis I diagnoses, were treated at this facility on
16 Innovations in CLINICAL NEUROSCIENCE [VOLUME 10, NUMBER 5–6, MAY–JUNE 2013]
average over three months, with TABLE 2. Changes in referral source diagnoses across during Years 2 and 3 of the study after
regular therapist and psychiatric application of structured diagnostic interviews
consultations and round-the-clock
clinical observation. REFERRAL DIAGNOSIS YEAR 2, % CONFIRMED YEAR 3, % CONFIRMED
Primary axis-I diagnoses were
examined during the entire period of
Bipolar 40 50
treatment for each case during the
three year period. Changes in the
Major depression 50 65
original admission diagnosis prior to
discharge from treatment were
recorded as the primary outcome Schizoaffective 10 11
measure. As a secondary outcome in
Years 2 and 3, the original clinician Schizophrenia 50 57
diagnosis was compared to the
admission diagnosis assigned Anxiety 75 50
following the structured diagnostic
interview. In the calculation of Overall confirmation rate 44 58
“change in diagnosis,” we used the
following rules: 1) We generated
global categories in order to avoid depression was changed to a with the structured procedures.
characterizing minor changes in diagnosis of bipolar disorder or vice These data are presented in Table 2.
diagnoses as discrepant. For instance, versa. In marked contrast, the rates See Figure 1 for a graphic depiction
we considered a diagnosis of bipolar of change of diagnoses generated by of these results. There were
II and bipolar I disorder to be the SCID during similarly lengthy substantial discrepancies between
consistent, although a change in stays was four percent and the rates these diagnoses. Of the cases
diagnosis from bipolar depression to of changes in MINI diagnoses was 11 interviewed with the SCID, 56
major depression was considered a percent. Chi-square tests were used percent of the cases were assigned a
change. 2) We did not consider to compare the differences in rate of diagnosis that was different from that
schizophrenia subtypes as part of the change in diagnoses across provided by the referral source and
diagnostic agreement, but considered procedures. The difference in rates of for the MINI the number of cases
schizophrenia to be different from change between the clinical whose diagnosis was different was 42
schizoaffective disorder. 3) Changes diagnostic assessment procedure and percent. Diagnostic confirmation
in clinical state codes (i.e., severe to the SCID was significant, rates for bipolar were 40 percent and
remission) during the course of chi2(1)=19.09, p<0.001, as was the 50 percent for the two years, and
treatment within the same diagnosis difference between clinical diagnoses confirmation of major depressive
were not considered as a difference and the MINI, chi2(1)=7.50, p<0.005. disorder were somewhat higher. Most
in diagnosis. However, the difference in rates of diagnostic discrepancies were
RESULTS changes in diagnoses between SCID interview-based diagnoses of major
As can be seen in Table 1, the and MINI procedures was not depression in cases referred as
overall pattern of admission significant, chi2(1)= 2.01, p=0.16. bipolar and bipolar depression in
diagnoses changed slightly over the In order to determine whether cases referred as major depression.
study period. Diagnoses of major there were differences across the Diagnostic confirmation of
depressive disorder became more three raters in the extent to which schizoaffective disorder was also very
common and diagnoses of psychosis their diagnoses were changed over low at 10 percent and 11 percent
became less so. When the primary the course of treatment for the across the two years.
outcome, change in diagnosis from patients, we performed a 3 (rater) x
admission to discharge, was 2 (changed or not changed) Chi- DISCUSSION
examined there were clear square test for each of the three Given the importance of matching
differences across the methods. In years. All three years suggested no appropriate diagnosis with evidence-
the year prior to the implementation differences across the three raters in based pharmacologic,
of the SCID, 74 percent of admission the extent to which their diagnoses psychotherapeutic, and psychosocial
diagnoses based on referral diagnosis were changed by the clinicians (all interventions, identifying an accurate
and unstructured interview were chi2(2 df)<1.47, all p>0.48). working diagnosis quickly and
changed over the course of the In the analyses of data from Years efficiently in community mental
treatment period, with the most 2 and 3, we compared clinical health settings is essential. Relying
common changes being that an diagnosis provided by the referring on referring diagnoses and self-
admission diagnosis of major source to the diagnoses generated report of previous diagnoses may
between referring clinician diagnoses
and the results of structured
assessments cannot be attributed to
bias on the part of clinicians toward
not modifying a diagnosis. The
facility at which these analyses
occurred is a private, non-profit
treatment program that does not bill
Medicaid or Medicare. A selection
bias in referred patients may limit
generalizability of these outcomes to
other treatment settings in which
patients from more varied economic
groups are assessed and the
applicability to inpatient settings
cannot be determined. The very
short stays typical in current
inpatient treatment make the use of
structured diagnostic assessments
FIGURE 1. Confirmation of referral source diagnoses across during Years 2 and 3 of the study
less useful. A relatively younger age
after application of structured diagnostic interviews of the patients in this study may limit
generalizability of the findings to
older patients who have had longer
yield a diagnosis that requires diagnosis of bipolar disorder in the experiences with serious, persistent
modification, even in the context of absence of symptomatic evidence in mental illnesses. Neither the SCID
an initial unstructured psychiatric patients eventually diagnosed with nor the MINI was designed for the
diagnostic assessment by experts. major depression and tendencies to purpose of diagnosing axis-II
Implementing structured interviews miss euphoric bipolar symptoms. pathology and as a result, these
in such settings may be prudent, as Unsystematic assessment may diagnoses could not be systematically
rates of diagnostic changes were produce both types of diagnostic assessed in this study. The required
significantly reduced in our study errors: inadequate knowledge of the investment in training clinicians and
using both the SCID and MINI. signs of a manic or hypomanic administering the standardized
Length of administration of the MINI episode may lead to a false positive assessments may similarly limit the
is shorter than administering the bipolar diagnosis and failure to practicality of assessment of these
SCID (20 minutes compared to over assess for manic episodes may lead outcomes in busy community mental
90 minutes), and diagnostic stability to false negatives. Schizoaffective health treatment centers, where
was not notably different between disorder seems to be more often resources tend to be limited.
these two structured interviews in found in the diagnostic opinions of Although the study participants
our cohorts. clinicians than in the results of were on average quite young,
Themes of modifications in structured assessments. patients were typically not recovering
referral diagnoses in our study There are some limitations to from their first episode of illness,
mirrored data from other community these data and these analyses. during which time actual changes in
diagnostic studies. In our Year-2 and Stability is not the only element of symptomatology and presentation
Year-3 cohorts, patients who validity; we did not examine might account for diagnostic
presented initially with diagnoses of treatment response, biomarkers, or uncertainty. As treatment options for
bipolar illness, unipolar depression, course of illness as validity mental illnesses continue to improve,
and schizoaffective disorder had indicators. We could not quantify the diagnostic stability and reliability
significant rates of re-diagnosis of reasons why clinicians changed become even more important in
their primary mental illness using a admission diagnoses, and there may community mental health settings.
structured diagnostic tool. Bipolar be several reasons for these changes. Pharmacologic, psychotherapeutic,
illness prior to admission seemed to It also is possible that clinicians were and social interventions used to treat
be both over-diagnosed and under- less likely to change diagnoses patients with bipolar illness are
recognized within patients referred generated by structured interviews, significantly different than similar
to this treatment facility. While the but the clinicians were not aware we classes of interventions for patients
origin of the pattern of such planned to examine changes in with personality disorders or even
diagnostic changes is unclear, there diagnosis as an outcome variable in a unipolar depression. This study
was both a bias toward over- research study. Discrepancies suggests that the up-front investment
18 Innovations in CLINICAL NEUROSCIENCE [VOLUME 10, NUMBER 5–6, MAY–JUNE 2013]
of effort and time to use a structured Psychiatr Res. 2007; 41:1012–1018. Disorder. Washington, DC: American
diagnostic assessment at the time of 7. Burdick KE, Endick CJ, Goldberg JF. Psychiatric Press, Inc.;1999:1–22.
admission to residential, partial Assessing cognitive deficits in bipolar 19. Jäger M, Haack S, Becker T, Frasch
hospitalization, and intensive out- disorder: are self-reports valid? K. Schizoaffective disorder: an
patient programs may be a wise Psychiatry Res. 2005;136:43–50. ongoing challenge for psychiatric
course of action for patients and 8. Harvey PD, Heaton RK, Carpenter, nosology. Eur Psychiatry.
payers alike. Matching specific and WT, Jr., et al. Diagnosis of 2011;26:159–165.
timely treatment to the appropriate schizophrenia: consistency across 20. Malhi GS, Green M, Fagiolini A, et al.
diagnosis makes sense for all information sources and stability of Schizoaffective disorder: diagnostic
stakeholders, despite the requisite the condition. Schizophr Res. issues and future recommendations.
time involved in administering the 2012;140:9–14. Bipolar Disord. 2008;10:215–230.
assessments. 9. Schwartz JE, Fennig S, Tanenberg- 21. Lake CR, Hurwitz N. Schizoaffective
Future directions for efficiently Karant M, et al. Congruence of disorder merges schizophrenia and
diagnosing mental illness in diagnoses 2 years after a first- bipolar disorders as one disease:
community psychiatric facilities admission diagnosis of psychosis. there is no schizoaffective disorder.
should include assessing variables Arch Gen Psychiatry. Curr Opin Psychiatry.
related to patients’ socioeconomic 2000;57:593–600. 2007;20:365–379.
factors, referral sources, age of 10. Bromet EJ, Naz B, Fochtmann LJ, et 22. First MB, Spitzer RL, Gibbon M, et
patients, and stigma. Although al. Long-term diagnostic stability and al. User’s guide for the Structured
currently limited in application, outcome in recent first-episode Clinical Interview for DSM-IV Axis
including biomarkers and imaging cohort studies of schizophrenia. I (SCID-I). Washington, DC:
data to make clinical diagnoses will Schizophr Bull. 2005;31:639–664. American Psychiatric Press,
also help determine cost-effective 11. Yutzy SH, Woofter CR, Abbott CC, et Inc.;1995.
Harnett-Sheehan K, et al. The Mini
and practical structured diagnostic al. The increasing frequency of 23. Sheehan DV, Lecrubier Y,
tools that busy community clinicians mania and bipolar disorder: causes
may implement in their treatment and potential negative impacts. J International Neuropsychiatric
planning for patients. Nerv Ment Dis. 2012;200:380–387. Interview (M.I.N.I.): the
12. Keck PE Jr, Kessler RC, Ross R. development and validation of a
REFERENCES Clinical and economic effects of structured diagnostic psychiatric
1998;59(Suppl 20):22–33.
1. Beck AT, Ward CH, Mendelson M, et unrecognized or inadequately interview. J Clin Psychiatry.
al. Reliability of psychiatric diagnosis treated bipolar disorder J Psychiatr
2: a study of consistency of clinical Pract. 2008;14(Suppl 2):31–38.
judgments and ratings. Am J 13. Hoch P, Pollatin P. Pseudoneurotic
Psychiatry. 1962;119:351–357. forms of schizophrenia. Psychiatric
2. Wing JK, Birley JL, Cooper JE, et al. Q. 1949;23:248–276.
Reliability of a procedure for 14. Pogge DL, Wayland-Smith D,
measuring and classifying “present Zaccario M, et al. Diagnosis of manic
psychiatric state. Br J Psychiatry. episodes in adolescent inpatients:
1967;113:499–515. structured diagnostic procedures
3. Spitzer RL, Endicott J. Current and compared to clinical chart diagnoses.
Past Psychopathology Scales. New Psychiatry Res. 2001;101:47–54.
York: New York State Psychiatric 15. Zimmerman M, Ruggero CJ,
Institute, Biometrics Research Chelminski I, Young D. Is bipolar
Division; 1968. disorder overdiagnosed? J Clin
4. American Psychiatric Association. Psychiatry. 2008;69:935–940.
Diagnostic and Statistical Manual 16. Hirschfeld RM, Vornik LA.
of Mental Disorders, 3rd edition. Recognition and diagnosis of bipolar
Washington, DC: American disorder. J Clin Psychiatry.
Psychiatric Press, Inc.; 1980. 2004;65(Suppl 15):5–9.
5. Amador XF, Flaum M, Andreasen 17. Leboyer M, Kupfer DJ. Bipolar
NC, et al. Awareness of illness in disorder: new perspectives in health
schizophrenia and schizoaffective care and prevention. J Clin
and mood disorders. Arch Gen Psychiatry. 2010;71:1689–1695.
Psychiatry. 1994;51:826–836. 18. Harvey PD, Yehuda R. Strategies to
6. Bowie CR, Twamley EW, Anderson study risk for Post-traumatic stress
H, et al. Self-assessment of disorder. In Yehuda R (ed). Risk
functional status in schizophrenia. J Factors for Post-traumatic Stress
Copyright of Innovations in Clinical Neuroscience is the property of Matrix Medical
Communications, LLC and its content may not be copied or emailed to multiple sites or
posted to a listserv without the copyright holder’s express written permission. However, users
may print, download, or email articles for individual use.
1.What was the purpose of this study?
2.Who were the participants in this study? (Give as much detail as possible.)
3.Who was responsible for assessing the patients in this study and what did they use to assess the patients?
4.According to the article, what are some of the pros and cons of structured interviews?
5.According to the article, why is correct, early diagnosis important?
6.What patients were excluded from this study and why?
7.Why was inter-rater reliability not a significant consideration in this study?
8.Based on the results of the study, which diagnostic tool generated the most reliable/stable diagnosis?
9.Based on the results of the study, what do the researchers suggest about using structured interviews for diagnosis?
TO GET YOUR ASSIGNMENTS DONE AT A CHEAPER PRICE,PLACE YOUR ORDER WITH US NOW