Volume 73, No.12: 2021 Siriraj Medical Journal
https://he02.tci-thaijo.org/index.php/sirirajmedj/index
793
Original Article
SMJ
Kantanut Yutrirak, M.D.*, Woraphat Ratta-apha, M.D., Ph.D.*, Pittaya Dankulchai, M.D.**, Panate Pukrittayakamee, M.D.*
*Department of Psychiatry, **Division of Radiation Oncology, Department of Radiology, Faculty of Medicine Siriraj Hospital, Mahidol University,
Bangkok 10700, ailand.
Psychometric Properties of the PHQ-9, HADS,
and CES-D Questionnaires and the Prevalence of
Depression in Patients with Cancer Receiving
Radiotherapy
ABSTRACT
Objective: e primary aim was to compare the psychometric properties among the Patient Health Questionnaire
(PHQ-9) (both including and excluding somatic symptom items), the depression subscale of the Hospital Anxiety
and Depression Scale (HADS-D), and the Center for Epidemiologic Studies Depression Scale (CES-D) in detecting
depression in cancer patients receiving radiotherapy. e secondary aim was to investigate the prevalence of
depression in this group of patients.
Materials and Methods: Overall, 198 participants with cancer diagnosis from a radiotherapy clinic took part in
the study. ey completed PHQ-9, HADS-D, and CES-D questionnaires and were interviewed in line with the
Mini-International Neuropsychiatric Interview (M.I.N.I.) to conrm the diagnosis. e PHQ-9 was analyzed for
three scoring methods: sum-score, inclusive (including all items), and exclusive (excluding 4 somatic symptom
items) methods. e psychometric properties of each questionnaire were analyzed. e prevalence of depression
measured by the M.I.N.I. was evaluated.
Results: e sum-score method of the PHQ-9 had an equal sensitivity (100%) to the HADS-D and CES-D, and had
a slightly higher specicity (91.1%) than the HADS-D (87.4%) and CES-D (90.6%). When compared results within
the PHQ-9, the sum-score method had greater sensitivity than the inclusive (71.4%) and exclusive (42.9%) methods,
and had a slightly lower specicity than the inclusive (96.9%) and exclusive (97.4%) methods. e prevalence of
depression assessed by the M.I.N.I was 3.5%.
Conclusion: e sum-score method of the PHQ-9 seemed to be the best tool to use for depression screening in
cancer patients receiving radiotherapy due to its excellent sensitivity and specicity.
Keywords: PHQ-9; HADS; CES-D; Depression; Cancer; Radiotherapy (Siriraj Med J 2021; 73: 793-800)
Corresponding author: Panate Pukrittayakamee
E-mail: panate.puk@mahidol.ac.th
Received 8 April 2021 Revised 8 September 2021 Accepted 5 October 2021
ORCID ID: https://orcid.org/0000-0001-8776-2427
http://dx.doi.org/10.33192/Smj.2021.103
INTRODUCTION
Depression is a common problem in patients with
cancer. In one meta-analysis, the prevalence of depression
among cancer patients was found to be 14.9%.
1
It has
been reported that depression increases the mortality
rate
2
, decreases the quality of life
3
, and decreases the will
to live of patients with cancer.
4,5
So, eective screening
for depression is required among patients with cancer.
e depression screening tools commonly used in
patients with cancer include the Patient Health Questionnaire
(PHQ-9)
6
, Hospital Anxiety and Depression Scale (HADS)
7
,
and Center for Epidemiologic Studies Depression Scale
Volume 73, No.12: 2021 Siriraj Medical Journal
https://he02.tci-thaijo.org/index.php/sirirajmedj/index
794
(CES-D).
8
e PHQ-9 was developed based on the major
depressive episode criteria of the Diagnostic and Statistical
Manual of Mental Disorders (DSM IV-TR). It is a 9-item
self-report questionnaire, which can be scored using a
sum-score method or a DSM IV-TR-based algorithm.
It has shown satisfactory concurrent and discriminant
validity and also reliability when validated in patients with
cancer.
9
e HADS is also a self-report questionnaire,
consisting of 14 items divided into depression and anxiety
subscales. It was developed for screening depression
and anxiety in a general medical population. Validation
studies of the HADS in cancer patients showed it had a
stable factor structure, moderate to high discriminant
validity, and adequate internal consistency.
9
e CES-D
is a 20-item self-report questionnaire developed for
screening depression in a general medical population and
in patients with cancer.
10
Results from validation studies
in cancer patients demonstrated its good sensitivity,
specicity, and internal consistency.
9
Although all these
three self-report questionnaires are easy to complete by
patients with physical illnesses and have been validated
in cancer populations, there is no consensus on which
screening tool is preferred for screening depression in
cancer patients.
Screening as well as diagnosing depression in patients
with cancer is challenging as cancer can produce somatic
symptoms that are similar to somatic symptoms of
depression, such as a decreased appetite, weight loss,
sleep problems, and fatigue.
11
Suggestions have been
made to exclude these somatic symptoms when evaluating
depression in cancer patients. Indeed, a previous study of
the PHQ-9 tried to explore the eect of excluding somatic
symptom items on detecting depression. In that study,
4 somatic symptom items, namely decreased appetite,
sleep problems, fatigue, and psychomotor retardation,
were excluded from the questionnaire and depression
was diagnosed when 3 of the remaining 5 items were
present. e results demonstrated that excluding those
items when assessing somatic symptoms of depression
had very little eect on detecting depression.
12
However,
the limitation of that study was that the gold standard
used for validity testing was not a structured diagnostic
interview.
Since there is insucient evidence for making a
recommendation about which depression screening tool
should be used in patients with cancer, this study
aimed to
compare the psychometric properties of the PHQ-9, HADS,
and CES-D in detecting depression in cancer patients in a
radiotherapy clinic. is study focused on cancer patients
in a radiotherapy clinic because these patients represent
variations in cancer type and stage. Furthermore, evidence
regarding the eect of excluding somatic symptom items
from the PHQ-9 remains inconclusive due to the lack
of using a diagnostic interview as the gold standard in
validity testing. Hence, this study also aimed to compare
the psychometric properties of the PHQ-9 between
including and excluding somatic symptom items by using
a structured diagnostic interview as a gold standard.
Finally, this study aimed to investigate the prevalence
of depression in cancer patients receiving radiotherapy.
MATERIALS AND METHODS
Participants
Cancer patients with any type and any stage of cancer
receiving treatment in a radiotherapy clinic of a tertiary
care hospital were recruited from January to April 2020.
e calculated sample size was 195. is sample size was
calculated by using the Wayne formula and based on a
prevalence of depression of 14.9% in cancer patients.
1
Measures
e ai version of the PHQ-9, the depression
subscale of the HADS (HADS-D), and CES-D were
used to assess depression. e ai version of the Mini-
International Neuropsychiatric Interview (M.I.N.I.) was
used as the gold standard to conrm a diagnosis of major
depressive episodes.
1) PHQ-9
e PHQ-9 is a 9-item self-report questionnaire
which can be scored using two methods: a sum-
score method with a cut-o score and an algorithm
scoring method. In the sum-score method, each
item can be rated from 0 to 3, with the total score
ranging from 0 to 27. Patients are classied
as having depression when the total score of the
ai version of the PHQ-9 is 9 or more.
13
However,
the cut-o score used in this study was re-calculated
to nd the most appropriate cut-o score for
cancer patients in this study. In the algorithm
scoring method, each item of the PHQ-9 is
counted as meeting a criterion if it is rated as 2
or 3. Patients are classied as having depression
when 5 of 9 items meet the criteria, one of which
must be item 1 (loss of pleasure in doing things)
or 2 (depressed mood).
6,14
e algorithm scoring
method in this study was split into two sub-
methods for analysis: an inclusive and exclusive
method. In the inclusive method, all 9 items
of the PHQ-9 were included in the assessment. In
the exclusive method, 4 items assessing somatic
symptoms of depression were excluded in order
to prevent false-positive results. ese items
Yutrirak et al.
Volume 73, No.12: 2021 Siriraj Medical Journal
https://he02.tci-thaijo.org/index.php/sirirajmedj/index
795
Original Article
SMJ
were item 3 (sleep problems), 4 (fatigue), 5
(appetite changed), and 8 (psychomotor retardation).
Patients were classied as having depression
when 3 of the remaining 5 items met the criteria,
one of which must be item 1 or 2.
11,12,15
2) HADS
e HADS is a 14-item self-report questionnaire,
with 7 items for the anxiety subscale and 7
items for the depression subscale.
7
However,
only the depression subscale of HADS (HADS-D)
was used in this study. For each subscale, each
item can be rated from 0 to 3, with the total
score ranging from 0 to 21. A sum score of 11
or more in the ai version of the HADS represents
depression.
16
However, the cut-o score used in
this study was re-calculated, as was also done
with the PHQ-9.
3) CES-D
e CES-D is a 20-item self-report questionnaire.
Each item can be rated from 0 to 3, with the
total score ranging from 0 to 60.
8
A sum score of
19 or more in the ai version of the CES-D
represents depression.
17
However, the cut-o
score used in this study was re-calculated, as
was also done with the PHQ-9 and the HADS.
4) M.I.N.I.
e ai version of the M.I.N.I. was translated from
the M.I.N.I. 5.0.0/DSM-IV. It is a structured
diagnostic interview comprising 16 modules for
assessing common psychiatric disorders. In this
study, the major depressive episode module was
used as the diagnostic tool. is module had a
sensitivity of 98% and specicity of 94%.
18
Data collection
Ethics approval was obtained from the Siriraj
Institutional Review Board. All the participants completed
the demographic data, PHQ-9, HADS-D, and CES-D
questionnaires. ey were interviewed using the M.I.N.I.
either by a psychiatric resident or a psychologist who had
been trained and certied in M.I.N.I.. Both interviewers
were blinded from the result of the self-rated questionnaires.
If depression was conrmed by M.I.N.I., the interviewers
would notify the attending physician to consider referring
the participant to consult psychiatrist for evaluation
and proper treatment. Data about cancer type, stage,
treatment, pain score, and opioid use were obtained
from the patients’ medical records.
Statistical analysis
e analysis was done with SPSS version 24. By
using the M.I.N.I. as the gold standard, the cut-o scores
of the PHQ-9, HADS-D, and CES-D were determined
by plotting their receiver operating characteristic (ROC)
curves. e psychometric properties of each questionnaire
were analyzed and demonstrated in terms of sensitivity,
specicity, positive predictive value (PPV), negative
predictive value (NPV), and likelihood ratio. Convergent
validity between the PHQ-9 and HADS-D, between PHQ-
9 and CED-D, and between HADS-D and CES-D were
analyzed by Spearman’s rho correlation. e internal
consistency of each questionnaire was analyzed by
Cronbach’s alpha. e prevalence of depression measured
by each questionnaire and the M.I.N.I. were evaluated.
RESULTS
In total, 198 participants were enrolled on the study,
and their demographic data are shown in Table 1. ere
was nearly an equal number of female (53.3%) and male
(46.5%) participants. Half the participants were more than
60 years old. e majority of participants (86.9%) were
recruited from an outpatient clinic. e most common
cancer types were breast (27.3%), prostate (13.6%), oro-
pharyngo-laryngeal (12.1%), and gastrointestinal cancer
(10.1%). e most common stage was the non-metastatic
stage (83.8%). Most of the participants did not have pain
(62.6%) and did not receive opioids (91.9%).
e most appropriate cut-o scores were 11 for
the PHQ-9, 7 for the HADS-D, and 20 for the CES-D.
e ROC curves for these cut-o values are displayed
in Fig 1. e area under the curve values for each were
0.97 (SD = 0.01; 95% CI 0.94 to 0.99) for the PHQ-9,
0.95 (SD = 0.02; 95% CI 0.91 to 0.98) for the HADS-D,
and 0.98 (SD = 0.01; 95% CI 0.95 to 1.00) for the CES-D.
All of these values show high accuracy.
19
The psychometric properties of the PHQ-9,
HADS-D, and CES-D are listed in Table 2. e sum-
score method used for the PHQ-9, the HADS-D, and
the CES-D demonstrated good sensitivity (100%) and
good specicity (91.1%, 87.4%, and 90.6%, respectively).
Although the inclusive and exclusive methods of the
PHQ-9 demonstrated slightly higher specicity than the
sum-score method (96.9% for the inclusive method and
97.4% for the exclusive method), their sensitivities were
much lower (71.4% for the inclusive method and 42.9%
for the exclusive method). Comparing the inclusive and
exclusive method, the inclusive method demonstrated
greater sensitivity with similar specicity. Convergent
validity testing showed good correlations between the
PHQ-9 and HADS-D (r = 0.67, p < 0.01), between the
Volume 73, No.12: 2021 Siriraj Medical Journal
https://he02.tci-thaijo.org/index.php/sirirajmedj/index
796
TABLE 1. Demographic data.
Characteristics (n = 198) n (%)
Gender
Female 106 (53.5)
Male 92 (46.5)
Age (mean 59.4, SD 13.3)
Education
High school or less 119 (60.1)
Undergraduate degree or more 79 (39.9)
Setting
Outpatient 172 (86.9)
Inpatient 26 (13.1)
Cancer type
Breast 54 (27.3)
Prostate 27 (13.6)
Oro-pharyngo-laryngeal 24 (12.1)
Gastrointestinal 20 (10.1)
Gynecologic 19 (9.6)
Lung 16 (8.1)
Brain 14 (7.1)
Others* 24 (12.0)
Disease stage
Non-metastasis 166 (83.8)
Metastasis 32 (16.2)
Treatment
Radiotherapy 16 (8.1)
Radiotherapy + Surgery 59 (29.8)
Radiotherapy + Chemotherapy 42 (21.2)
Radiotherapy + Surgery + Chemotherapy 81 (40.9)
Pain (mean 1.76, SD 2.8)
No pain 124 (62.6)
Mild (Pain score 1-3) 30 (15.2)
Moderate (Pain score 4-6) 23 (11.6)
Severe (Pain score 7-10) 21 (10.6)
Opioids use
No 182 (91.9)
Yes 16 (8.1)
*yroid 7, Hematologic 7, Liver 2, Urinary tract 3, Anus 1, Cholangiocarcinoma 1, Nasal cavity 1, Epithelioid tumor 1, Multiple
primary 1.
Yutrirak et al.
Volume 73, No.12: 2021 Siriraj Medical Journal
https://he02.tci-thaijo.org/index.php/sirirajmedj/index
797
Original Article
SMJ
TABLE 2. Psychometric properties.
Sensitivity Specicity PPV NPV + likelihood - likelihood Internal
consistency
(Cronbach’s alpha)
PHQ-9
Sum-score 100 91.1 29.2 100 11.2 0 0.804
(cut-off 11)
Algorithm scoring
Inclusive 71.4 96.9 45.5 98.9 22.7 0.3 -
Exclusive 42.9 97.4 37.5 97.9 16.4 0.6 -
HADS-D (cut-off 7) 100 87.4 22.6 100 8 0 0.772
CES-D (cut-off 20) 100 90.6 28 100 10.6 0 0.815
PHQ-9 and CES-D (r = 0.68, p < 0.01), and between
HADS-D and CES-D (r = 0.74, p < 0.01). e internal
consistencies of the PHQ-9 and CES-D were good
(Cronbach’s alpha = 0.80 and 0.82, respectively), while
the internal consistency of the HADS-D was acceptable
(Cronbach’s alpha = 0.77).
20
e prevalence of depression measured by each
questionnaire and the M.I.N.I. are presented in Table 3.
e prevalence measured by the inclusive method (5.6%)
and exclusive method (4.0%) of the PHQ-9 were close to
the prevalence measured by the M.I.N.I. (3.5%), which
represents the gold standard. However, the prevalence
measured by the sum-score methods of the PHQ-9
(12.1%), HADS-D (15.7%), and CES-D (12.6%) were
much higher than the prevalence measured by the M.I.N.I..
DISCUSSION
e study aimed to test the psychometric properties
of the three self-rating questionnaires PHQ-9, HADS-D,
and CES-D for screening depression in cancer patients
receiving radiotherapy. e results showed that the
psychometric properties, both validity and reliability, of
all questionnaires were nearly equivalent. Comparing
the sum-score methods of the PHQ-9, HASD-D, and
CES-D, the sum-score method of PHQ-9 is recommended
for depression screening in cancer patients receiving
radiotherapy because it showed high sensitivity and
the highest specicity and all of its items are similar to
the major depressive disorder criteria of the DSM-5.
14
In addition, because the PHQ-9 consists of 9 items that
can be completed within a few minutes, it is convenient
Fig 1. ROC curve
Volume 73, No.12: 2021 Siriraj Medical Journal
https://he02.tci-thaijo.org/index.php/sirirajmedj/index
798
for patients with physical illness. Although the CES-D
showed a similar specicity to the sum-score method
of the PHQ-9, the major limitation of the CES-D is it is
time consuming to complete because it consists of 20
items.
Regarding the PHQ-9, its sum-score method
demonstrated a much higher sensitivity but with a similar
specicity when compared to the algorithm scoring
methods. is nding suggested that the sum-score
method is better than the algorithm scoring methods for
screening depression in patients with cancer. Comparing
the methods within the algorithm scoring methods, the
exclusive method had a much lower sensitivity than the
inclusive method. is result reected that the items
concerning the somatic symptoms of depression should
not be excluded from the PHQ-9 when screening for
depression in cancer patients. is nding is supported
by evidence from another study which demonstrated
that the somatic symptoms of depression were more
likely to be present in depressed than in non-depressed
cancer patients.
21
The cut-off scores of the screening tools were
dierent from the recommendation from the previous
study. According to our ndings, the cut-o scores of
the PHQ-9, HADS-D, and CES-D were 11, 7 and 20,
respectively, while the cut-o scores of the ai version
recommended in previous studies were 9, 11, and 19,
respectively.
13,16,17
One of the reasons for this disparity may
be the dierence in somatic symptoms in the population
between the studies. Previous studies of the ai version
of the PHQ-9 was conducted in family medicine clinic
and the CES-D was conducted in general populations
which tend to have less somatic symptoms. While the
ai HADS-D study was conducted in in-patients with
cancer which tend to have more somatic symptoms.
To the best of our knowledge, the present study is the
rst one to investigate the cut-o score in this specic
population.
The prevalence of depression assessed by the
structured interview (M.I.N.I.) was nearly equal to the
prevalence assessed by the algorithm scoring methods of the
PHQ-9. is may be explained by the high specicity of
the algorithm scoring methods. Comparing the methods
within the algorithm scoring methods, the inclusive method
is preferred for assessing the prevalence or diagnostic
purpose due to its high specicity and higher sensitivity
than the exclusive method. However, inspection of the
raw data showed that some patients had a diagnostic
mismatch between the M.I.N.I. assessment and the
PHQ-9 algorithm scoring methods. us, evaluation of the
psychometric properties through a diagnostic interview
conducted by a psychiatrist as the gold standard should
be conducted in a further study in order to conrm
whether the PHQ-9 algorithm scoring methods are
appropriate for assessing the prevalence of depression
in patients with cancer.
In contrast, the prevalence as assessed by the PHQ-9
sum-score method, HADS-D, and CES-D was relatively
high when compared with the M.I.N.I. due to the false-
positive cases. Since these three questionnaires are scored
using a sum-score method, the severity ratings of the
somatic symptoms that overlap with cancer symptoms
need to be taken into account. As a consequence, cancer
symptoms may have an inuence on increasing the somatic
symptoms scores, leading to false-positive results.
12
We suggest that these three questionnaires may not be
appropriate for assessing the prevalence of depression
in patients with cancer.
Yutrirak et al.
TABLE 3. Prevalence.
%
M.I.N.I. 3.5
PHQ-9
Sum-score method 12.1
Algorithm scoring methods
Inclusive method 5.6
Exclusive method 4.0
HADS-D 15.7
CES-D 12.6
Volume 73, No.12: 2021 Siriraj Medical Journal
https://he02.tci-thaijo.org/index.php/sirirajmedj/index
799
Original Article
SMJ
e prevalence of depression assessed by the M.I.N.I.
in our study was lower than the average prevalence in
a meta-analysis in the literature (3.5% vs. 14.9%).
1
is
discrepancy may be due to the dierence in cancer stage
of participants among the studies. Our study and the
studies with a similar prevalence were conducted in
patients with cancer of any type and stage, mostly the
non-metastatic stage.
22-25
In contrast, the studies with a
prevalence of around 14.9% were conducted in cancer
patients within 12 months of diagnosis
26
, post-treatment
cancer patients
27
, and patients with recurrent or metastatic
cancer.
28
is may imply that patients are more likely to
develop depression when initially facing cancer diagnosis
and when facing advanced cancer. erefore, depression
screening should be performed within the rst year of
cancer diagnosis and upon progressing to an advanced
stage. Moreover, a systematic review reported that the
rate of depression is higher in adolescents and young
adults with cancer because of the disruptions in their
school life, career path, or early marital life.
29
It can be
implied that if we include more young age patients in
the study, we will gain more prevalence of depression.
Further study should be designed to include patients in
all age groups to improve the precision of the results.
Study limitations
Several limitations in the present study should be
considered. We use the M.I.N.I. as the gold standard
for depression diagnosis instead of using the standard
interview by psychiatrists because it consumed much less
time when must deal with the high volume of participants.
erefore, it could have some false positive and false
negative cases. e sensitivity and specicity in this
study may be dierent from a previous study conducted
in a population with a higher prevalence of depression.
30
Hence, further studies should be investigated in cancer
patients with a higher prevalence of depression, such as
newly diagnosed cancer patients, post-treatment cancer
patients, and patients with more advanced-stage cancer.
Furthermore, patients in a surgery and chemotherapy
clinic should be recruited to apply the results more
broadly.
CONCLUSION
e sum-score method of the PHQ-9 seemed to
be the best tool to use for depression screening among
cancer patients receiving radiotherapy. e inclusive
method of the PHQ-9 may be useful for prevalence
studies or could serve a diagnostic purpose due to its
high specicity and acceptable sensitivity. e prevalence
of depression assessed by the M.I.N.I. was 3.5%, nearly
equal to the prevalence assessed by the inclusive method
of the PHQ-9, which was 5.6%.
ACKNOWLEDGEMENTS
e authors would like to thank the Department of
Radiology for allowing us to collect the data. We would
like to thank Lakkhana ongchot, psychologist, for
helping collect the data and we also would like to thank
to Naratip Sanguanpanich, statistician, for statistical
analysis advice.
REFERENCES
1. Mitchell AJ, Chan M, Bhatti H, Halto M, Grassi L, Johansen C,
Meader N. Prevalence of depression, anxiety, and adjustment
disorder in oncological, haematological, and palliative-care
settings: a meta-analysis of 94 interview-based studies. Lancet
Oncol 2011;12:160-74.
2. Pinquart M, Duberstein PR. Depression and cancer mortality:
a metaanalysis. Psychological Medicine 2010;40:1797-810.
3. Smith HR. Depression in cancer patients: Pathogenesis, implications
and treatment (Review). Oncology Letter 2015;9:1509-14.
4. Chochinov HM, Tataryn D, Clinch JJ, Dudgeon D. Will to
live in the terminally ill. Lancet 1999;354:816-9.
5. Breirbart W, Rosenfeld B, Pessin H, Kaim M, Funesti-Esch
J, Galietta M, Nelson CJ, Brescia R. Depression, hopelessness,
and desire for hastened death in terminally ill patients with
cancer. JAMA 2000;284(22):2907-11.
6. Spitzer RL, Kroenke K, Williams JB. Validation and utility of
a self-report version of PRIME-MD: the PHQ primary care
study. JAMA 1999;282(18):1737-44.
7. Zigmond AS, Snaith RP. e Hospital Anxiety and Depression
Scale. Acta Psychiatr Scand 1983;67:361-70.
8. Radlo LS. e CES-D scale: a self-report depression scale for
research in the general population. Applied Psychological
Measurement 1977;1:385-401.
9. Vodermaier A, Linden W, Siu C. Screening for emotional distress
in cancer patients: a systematic review of assessment instruments.
J Natl Cancer Inst. 2009;101(21):1464-88.
10. Chhabria KS, Carnaby GD. Psychometric validation of the
Center for Epidemiological Studies Depression Scale in Head
and Neck Cancer patients. Oral Oncology 2017; 75:158-62.
11. Zimmerman M, Galione JN, Chelminski I, McGlinchey JB,
Young D, Dalrymple K, Ruggero CJ, Witt CF. A simpler
denition of major depressive disorder. Psychological Medicine
2010;40:451-7.
12. Lie HC, Hjermstad MJ, Fayers P, Finset A, Kaasa S, Loge JH.
Depression in advanced cancer – Assessment challenges and
associations with disease load. Journal of Aective Disorders
2015;173:176-84.
13. Lotrakul M, Sumrithe S, Saipanish R. Reliability and validity
of the ai version of the PHQ-9. BMC Psychiatry 2008;8:46.
14. American Psychiatric Association. Diagnostic and statistical
manual of mental disorder. 5
th
ed. Arlington, VA: American
Psychiatric Association; 2013.
15. Zimmerman M, Emmert-Aronson BO, Brown TA. Concordance
between a simpler denition of major depressive disorder and
diagnostic and statistical manual of mental disorders, fourth edition:
Volume 73, No.12: 2021 Siriraj Medical Journal
https://he02.tci-thaijo.org/index.php/sirirajmedj/index
800
an independent replication in an outpatient sample. Comprehensive
Psychiatry 2011;52:261-4.
16. Nilchaikovit T, Lortrakul M, Phisansuthideth U. Development
of ai version of Hospital Anxiety and Depression Scale in
cancer patients. J Psychiatr Assoc ailand 1996;41(1):18-30.
17. Kuptniratsaikul V, Pekuman P. e study of the Center for
Epidemiologic Studies-Depression Scale (CES-D) in ai
people. Siriraj Hosp Gaz 1997;49:442-8.
18. Kittirattanapaiboon P, Khamwongpin M. e Validity of
the Mini International Neuropsychiatric Interview (M.I.N.I.)-
ai Version. Journal of Mental Health of ailand 2005;13
(3)c.
19. Streiner DL, Cairney J. What’s under the ROC? An introduction to
receiver operating characteristics curves. Can J Psychiatry 2007;
52(2):121-8.
20. George D, Mallery P. SPSS for Windows step by step: A simple
guide and reference. 11.0 update (4
th
ed.). Boston: Allyn &
Bacon; 2003.
21. Mitchell AJ, Lord K, Symonds P. Which symptoms are indicative
of DSMIV depression in cancer settings? An analysis of the
diagnostic signicance of somatic and non-somatic symptoms.
J Aect Disord 2012;138(1-2):137-48.
22. Hardman A, Maguire P, Crowther D. The recognition of
psychiatric morbidity on a medical oncology ward. J Psychosom
Res 1989;33:235-39.
23. Derogatis LR, Morrow GR, Fetting J, Penman D, Piasetsky S,
Schmale AM, Henrichs M, Carnicke CL Jr. e prevalence of
psychiatric disorders among cancer patients. JAMA 1983;249:
751-57.
24. Kugaya A, Akechi T, Okuyama T, Okamura H, Uchitomi Y.
Screening for psychological distress in Japanese cancer patients,
Jpn J Clin Oncol 1998;28:333-38.
25. Lansky SB, List MA, Herrmann CA, Ets-Hokin EG, DasGupta
TK, Wilbanks GD, Hendrickson FR. Absence of major depressive
disorder in female cancer patients. J Clin Oncol 1985;3:1553-
60.
26. Akechi T, Okuyama T, Sugawara Y, Nakano T, Shima Y,
Uchitomi Y. Major depression, adjustments, and post-traumatic
stress disorder in terminally ill cancer patients: associated and
predictive factors. J Clin Oncol 2004;22:1957-65.
27. Gandubert C, Carrière I, Escot C, Soulier M, Hermès A, Boulet
P, Ritchie K, Chaudieu I. Onset and relapse of psychiatric disorders
following early breast cancer: a case–control study. Psychooncology
2009;18:1029-37.
28. Razavi D, Delvaux N, Farvacques C, Robaye E. Screening for
adjustment disorders and major depressive disorders in cancer
in-patients. Br J Psychiatry 1990;156:79-83.
29. Park EM, Rosenstein DL. Depression in adolescents and young
adults with cancer. Dialogues Clin Neurosci 2015;17(2):171-80.
30. Leeang MM, Rutjes AW, Reitsma JB, Hoo L, Bossuyt PM.
Variation of a test’s sensitivity and specicity with disease
prevalence. CMAJ 2013;185(11):E537-E544.
Yutrirak et al.
