-
Feb 6 1
Title: Reliability and validity of the Center for Epidemiologic
Studies- Depression scale in 1 screening for depression among HIV
infected and uninfected pregnant women attending 2 antenatal
services in northern Uganda 3 4 Authors: Barnabas K. Natamba1, 2*,
Thomas O. Oyok3, Angela Arbach4, Jane Achan5, Shibani 5 Ghosh6,
Saurabh Mehta1, Rebecca J. Stoltzfus1, Jeffrey K. Griffiths6, 7,
Sera L. Young1 6 7 Author affiliations: 1Division of Nutritional
Sciences, Cornell University, Ithaca, NY 14850, 8 USA. 2Department
of Public Health, Faculty of Medicine, Gulu University. Gulu.
Uganda. 9 3Department of Mental Health, Faculty of Medicine, Gulu
University. Gulu. Uganda. 4Weill 10 Cornell Medical College. New
York. NY 10021. USA. 5Department of Pediatrics and Child 11 Health,
Makerere University. Kampala. Uganda. 6Friedman School of Nutrition
Science and 12 Policy. School of Medicine, Veterinary Medicine,
Nutrition and Engineering. Tufts University, 13 Boston, MA 02155.
USA. 7Department of Public Health and Community Medicine. School of
14 Medicine, Veterinary Medicine, Nutrition and Engineering. Tufts
University. Boston. MA 02155. 15 USA. 16 17 *Corresponding author:
Division of Nutritional Sciences, Cornell University, Ithaca. N.Y.
18 14850. Email [email protected]; Fax: 1-607-255-1033 ; Tel:
+256782415501 19 20 21 22 23 24 25
-
Feb 6 2
Abstract 26 Introduction: Given the paucity of data on the
burden and measurement of depression in 27 pregnant women living in
high HIV burden resource poor settings, we aimed at determining the
28 reliability and validity of the Center for Epidemiologic Studies
Depression (CES-D) scale as a 29 screening tool for depression
among pregnant Ugandan women of mixed HIV status. 30 Methods: 123
(36 HIV-infected and 87 HIV-uninfected) pregnant women receiving
antenatal 31 care at Gulu Regional Referral Hospital were recruited
in their first and second trimesters and 32 underwent two
consecutive interviews: (1) a screening interview using the CES-D
scale and (2) 33 a diagnostic Mini-International Neuropsychiatric
Interview (MINI) for current major depressive 34 disorder (MDD)
performed by a psychiatrist. We employed measures of internal
consistency 35 (Cronbachs alpha), criterion validity (Area Under
the Receiver Operating Characteristic Curve 36 (AUROC), sensitivity
(Se) and positive predictive value (PPV)) and construct validity
(unadjusted 37 bivariate analyses between CES-D scores and maternal
HIV status, food insecurity score, and 38 tertile of per capita
household income) to evaluate the reliability and accuracy of the
CES-D 39 scale. 40 Results: 35.8% of the respondents were diagnosed
to be experiencing a current MDD as 41 defined by the MINI. The
CES-D had high internal consistency (Cronbachs alpha=0.92), 42
demonstrated good discriminatory ability in detecting MINI-defined
current MDDs 43 (AUROC=0.82) and had an optimum cut-off score for
screening for current MDDs of 15 (Se= 44 84.1%, PPV=77.2%). Lastly,
the CES-D showed strong construct validity with HIV infected 45
women, poorer women, or more food insecure women scoring higher on
this scale than their 46 better off counterparts. 47 Conclusions:
The CES-D is a suitable scale for screening for depression among
pregnant 48 women attending antenatal services in Uganda. 49 Key
words: pregnancy, major depression, prenatal depression, northern
Uganda 50
-
Feb 6 3
Introduction 51 In the two decades since the first Global Burden
of Disease (GBD) report was released 52
in 1990, the impacts that HIV infection and major depressive
disorders (MDDs) have had on 53 medical and public health systems
have changed profoundly. The 2010 GBD report indicates 54 that HIV
infection has risen from being the 33rd to the 5th contributor to
the global burden of 55 disease (Murray et al. 2013). At the same
time, the disease burden attributable to MDDs has 56 risen from
being the 15th to the 11th. Whereas there has been a surge in
information and 57 strategies for managing or preventing HIV
infection, practical strategies for managing MDDs are 58 still very
limited, particularly in resource poor settings. 59 60
In various settings in Sub-Saharan Africa, prevalence rates of
MDDs among adults living 61 with HIV ranging from 71.3% in Zambia
(Chishinga et al. 2011) to 47% in Uganda (Kaharuza et 62 al. 2006),
43.7% in South Africa (Myer et al. 2008), and 30% in Zimbabwe
(Chibanda et al. 63 2010) have been documented. A number of
biological factors, psychosocial variables, and 64 mental history
or comorbid psychiatric illnesses may help explain why persons with
HIV are at a 65 higher risk of depression (Arseniou et al. 2014).
Importantly, depression has been associated 66 with increased rates
of HIV disease progression and mortality (Antelman et al. 2007),
inequity in 67 decision-making and relationship power among
heterosexual couples (Hatcher et al. 2012), and 68 AIDS related
stigma (Simbayi et al. 2007). Whereas reports indicate that as many
as 30% of 69 persons living with HIV/AIDS develop depressive
disorders during the course of their illness 70 (Ciesla et al.
2001; Nakimuli-Mpungu et al. 2011), most studies reporting the
prevalence or 71 incidence of depressive symptoms among people
living with HIV (PLHIV) have not compared 72 this to HIV negative
persons. In one small cross-sectional study, where the prevalence
of 73 depression in newly tested HIV positive pregnant women were
compared to that of HIV negative 74 women, authors did not find an
association between HIV infection status and depression 75
-
Feb 6 4
(Rochat et al. 2006). Understanding the prevalence and drivers
of MDDs among HIV infected 76 and uninfected persons need further
clarification so as to enable mental health providers and 77
planners in resource poor settings better target appropriate care
for those that are affected 78 most. 79 80
In addition to the importance of the interaction between HIV
infection and MDDs in areas 81 with high HIV burden, it is
important to recognize that the two diseases disproportionately
affect 82 women and most likely this happens during the perinatal
period (Arseniou et al. 2014; Bennet et 83 al. 2004). Research from
both resource-poor (Roberts et al. 2008; Olley et al. 2004) and 84
resource-rich (Hwu et al. 1996) settings indicate that women are
more likely than men to report 85 a high prevalence of depressive
symptoms. Biological, sociocultural and psychological factors 86
may explain why women are at greater risk of HIV infection (Wingood
& DiClemente 2000) 87 and/or depression (Nolen-Hoeksema 2006)
than men. 88 89
Given the coexistence of HIV infection and MDDs in SSA,
combining screening for 90 depression and HIV in primary care
settings targeting women may be an efficient means to use 91
limited resources, especially if there is potential for treatment
and/or referral following 92 administration of screening tests.
Indeed, in 2010, the WHO developed the mental health Gap 93 Action
Program (mhGAP) that recommends interventions to address mental
health problems in 94 resource poor settings with limited
specialist care for such problems (WHO 2010). Antenatal 95 care
(ANC) clinics in regional, district, and lower level primary health
care centers in resource-96 poor settings offering services to
prevent the vertical transmission of HIV may also provide entry 97
points for the identification, treatment and referral of pregnant
women suffering from depression. 98 99
-
Feb 6 5
In order to appropriately identify participants with or at risk
for depression, theres need 100 for using a reliable and valid
screening tool. A review of depression screening tools used in low
101 and middle income countries did not recommend any specific
instrument, as no single 102 depression screening tool was found to
be superior to others (Akena et al. 2012). The Center for 103
Epidemiologic Studies-Depression (CES-D) (Radloff, 1977) is an
attractive scale for screening 104 for depression because it takes
a short period of time to administer, can be administered after 105
only a short training period, and covers all of the common symptoms
of major depression, i.e. 106 depressive mood, feelings of guilt
and worthlessness, psychomotor retardation, loss of appetite, 107
and sleep disturbance. The CES-D assesses 20 items designed to
measure self-reported 108 depressive symptoms in the prior week.
Each of the 20 items is assigned a value of 0-3, with 109 four
items positively worded and reverse scored. The total score is
computed by adding each of 110 the 20 items, such that the range is
0 to 60. A score of >16 in the general US population 111
suggests the presence of a major depressive disorder. Elsewhere,
the cutoff will depend on the 112 population studied and the extent
to which each of the 20 items are prevalent in such a 113
population. Thus, it is important that the scale be validated
before it is used in a non-US 114 population. 115 116
Within sub-Saharan Africa, the CES-D has been validated in
Zambia (Chishinga et al. 117 2011) and South Africa (Myers et al.
2008) by comparing results of the CES-D to those obtained 118 using
the gold standard for depression assessment, the depression module
of the Mini-119 International Neuropsychiatric Interview (MINI).
The MINI is a short, structured diagnostic 120 interview that was
developed in 1990 by psychiatrists and clinicians in the United
States and 121 Europe for DSM-IV psychiatric disorders (Sheehan et
al. 2006). Although the CES-D has been 122 used among PLHIV in
Uganda (Kaharuza et al. 2006 & Nakasujja et al. 2010), it has
actually not 123 been validated there. Further, all of these
studies occurred among HIV-infected adults only, 124
-
Feb 6 6
making it difficult to compare findings from such populations to
those of HIV negative 125 populations. Therefore, we aimed at
assessing the reliability, construct and criterion validity of 126
the CES-D as a measure of depression in both HIV-infected and
uninfected pregnant women in 127 northern Uganda. 128 129 130
Materials and methods 131 132 Design, setting and participants
133
Data were collected between 10th October 2012 and 21st December
2012, within the 134 context of the Prenatal Nutrition and
Psychosocial Health Outcomes Study (PreNAPs), a 135 longitudinal
cohort study aimed at describing and comparing nutritional and
psychosocial 136 exposures and outcomes among HIV infected and
uninfected pregnant women attending ANC 137 services in northern
Uganda. 138 139
Participants were recruited from the antenatal care (ANC) clinic
of Gulu Regional 140 Referral Hospital (GRRH) in Gulu, northern
Uganda, a busy primary care clinic at which more 141 than 400
initial antenatal visits occur monthly. As at other public medical
facilities in Uganda, all 142 services at GRRH including
medications, care for pregnant women with HIV, etc. are offered 143
free of charge. 144 145
Both HIV infected and uninfected pregnant women presenting at
ANC between 10 and 146 26 weeks of gestation and who resided within
30 KM of GRRH were invited to participate. 147 Consistent with
Ugandan national policy, all HIV infected women were receiving
prophylactic 148 ARVs. Two women whose HIV status could not be
ascertained after they were interviewed were 149 excluded. HIV
infected women were oversampled to achieve a ratio of 1 HIV
infected: 2 HIV 150
-
Feb 6 7
uninfected participants, a proportion that is much higher than
the 10.3% age adjusted 151 prevalence of HIV observed at various
ANC clinics in northern Uganda (Fabian et al. 2007). 152 153 Data
collection and analysis 154
Before interviews began, all study instruments were translated
by local research staff 155 into Acholi and Langi, the two
predominant and similar Luo languages in the study communities. 156
The questionnaires were back-translated into English by the same
team and discrepancies in 157 conceptual and semantic equivalence
were resolved through discussion involving all the 158 translators,
the research assistants, and the GRRH psychiatrist. 159 160
The CES-D was administered by research assistants after a
baseline health and 161 demographics data were collected. Then, on
the same day, respondents underwent a diagnostic 162 interview by
the GRRH psychiatrist (TOO) using the MINI-depression module. The
MINI 163 diagnostic interview was conducted by a psychiatrist who
was blinded to the results of the CES-164 D screening interview.
165 166
Study participants were categorized into cases and non-cases of
depressed patients 167 based on the MINI outputs for currently
experiencing a MDD. Median total scores on the CES-D 168 were then
compared against the MINI-defined diagnosis of current MDD using
the Wilcoxon 169 rank-sum test. We used multiple methods to
determine the reliability (or internal consistence), 170 criterion
and construct validity of the CES-D. Firstly we assed the
reliability of the entire CES-D 171 scale using the Cronbachs test
for internal consistency. Cronbachs alpha coefficients of 0.7 or
172 greater were considered to be reliable (Santos 1999). Secondly,
we determined the CES-Ds 173 criterion validity i.e. the extent to
which measurements obtained using the CES-D concurred with 174
those determined with an established diagnostic test or gold
standard (Porta 2008 ), by 175
-
Feb 6 8
comparing scores on the CES-D with psychiatrist administered
MINI outputs. We used a non-176 parametric method, area under the
receiver operating characteristic curve (AUROC) analysis, to 177
estimate the diagnostic accuracy of the CES-D. The optimum cut-off
scores consistent with 178 current MDD were determined for this
population based on the psychiatric assessment. The cut-179 off
score that maximized sensitivity and positive predictive value
(PPV) was selected. 180 181
Lastly, we determined the construct validity, i.e. the extent to
which a given 182 measurement corresponds to theoretical concepts
(constructs) concerning the phenomena 183 under study (Porta 2008
& Straus 2009), by conducting bivariate analyses to determine
184 whether, as we had hypothesized, CES-D scores were higher for
HIV infected, poorer, or food 185 insecure pregnant women. Maternal
HIV status was determined at the ANC clinic following 186
Government of Uganda HIV counseling and testing guidelines and
prior to enrollment into this 187 study. Wealth was operationalized
based on household per capita income obtained by summing 188 the
estimated monthly income of all income earning household members
and dividing this figure 189 by the number of all household members
including the respondent. Given the non-normal 190 distribution of
per capita household income data, we used tertiles of household per
capita 191 income as the independent variable in the bivariate
model with CESD scores as the outcome 192 variable. Individual
level FI was assessed using the Individually-focused Food
Insecurity Access 193 Scale (IFIAS)- a modified version of the
Household Food Insecurity Access Scale (HFIAS; 194 Coates et al.
2007). Unlike the HFIAS, and given potential for intra-household
differences in the 195 experience of food insecurity, the IFIAS
assesses perceptions of inadequate access to food at 196 the level
of the individual. IFIAS scores and categories were derived using
the methods 197 recommended by Coates et al. (2007). 198 199
Data were analyzed using Stata 12 (College Station, Texas, USA).
200
-
Feb 6 9
201 Ethical considerations 202
The study protocol was approved by Cornell University
Institutional Review Board, Gulu 203 University Institutional
Review Committee and the Ugandan National Council for Science and
204 Technology. Written informed consent was obtained from all
study subjects before enrollment. 205 Patients identified to be
severely depressed, as per psychiatric assessment using the
MINI-206 Depression section, were referred for further assessment
and treatment at GRRH. 207 208 Results 209 210 Characteristics of
the study participants 211
One hundred and thirty pregnant women were approached to
participate in the MDD 212 sub-study of Pre-NAPs, and 125 (96.2%)
of them consented to participate. Nonresponse was 213 mainly due to
lack of time (n=5). Of the 125 participants who agreed to take part
in the study, 214 two did not complete the entire interview due to
lack of time, and the rest (36 HIV-infected and 215 87
HIV-uninfected) participated in both the CES-D and MINI. 216
217 The median age of study participants was 23 years, a
majority (87.8%) were married or 218
living in stable heterosexual relationships and spoke either
Acholi or Langi (94.3%). Generally, 219 study participants were of
low socioeconomic status, as 53.7% had only primary education or
220 less, 50.4% described their main occupation as housewives,
52.0% perceived themselves to be 221 severely food insecure and the
median monthly household income per capita was 50,000 222 Uganda
shillings (approximately 20 US dollars). Further, 13% (16
participants) mentioned not 223 having any household income. 224
225
-
Feb 6 10
Internal consistency and validity of the CES-D scale 226 CESD
scores ranged between 0 and 57 with a mean (SD) of 17.1 (13.8). On
the other 227 hand, 35.8% of participants qualified for MINI
diagnosed major depressive disorder. The internal 228 consistency
of the CES-D was high (Cronbachs alpha =0.92). As expected, CES-D
scores were 229 higher among MINI-defined current MDD cases than
non-MDD cases (Figure 1, P
-
Feb 6 11
Criterion and construct validity. The AUROC, a measure of the
accuracy of 251 discriminating between two groups, was 0.82 for
differentiating women with current MDDs from 252 those without
based on their CES-D score. Our result is similar to AUROC values
reported from 253 validation studies conducted among HIV infected
persons in South Africa (Myers et al 2008; 254 AUROC=0.76) and TB
and HIV patients in Zambia (Chishinga et al 2011; AUROC=0.78). Per
255 our expectations, we observed higher scores on CES-D by women
who were HIV infected, in 256 the lower tertile of per capita
household income, or perceived themselves to be more food 257
insecure (with higher IFIAS scores) than their better off
counterparts. 258
A highly sensitive test is needed for screening examinations in
routine clinical care to 259 identify potential cases, while a
highly specific test is preferred for confirmatory purposes. Good
260 screening tests are expected to minimize the number and
proportion of false positive results, 261 and as such have a high
positive predictive value, or PPV. A test with high PPV is in part
valued 262 because it reduces the expense and risks related to
further evaluation of false positive tests 263 especially in
settings with limited resources (Patel et al. 2008). In this study,
we found that the 264 CES-D with a cut-off score of 15 achieved the
goal of suitably high sensitivity and moderate to 265 high PPV for
the detection of MDD (Sensitivity 84.1%, PPV 77.2%). The cut-off
score of 15 is 266 comparable to the cut-off score of 16 suggested
by Radloff for the general US population 267 (Radloff 1977). In
lower lever primary care settings with limited access to
specialized mental 268 health workers capable of diagnosing
prenatal depression, a cut-off score of 15 that maxims 269 both
sensitivity and positive predictive value is recommended. However,
for higher-level hospital 270 settings with fully equipped mental
health departments and potential for referral to specialized 271
depression care, a CES-D score of 16 may be used during the
screening stage, provided the 272 resultant large number of false
positive results are reviewed by qualified mental health 273
providers. 274
275
-
Feb 6 12
Our study has some limitations. The generalizability of our
findings may be limited. First, 276 we oversampled HIV infected
women, and it is possible that the CES-D performs differently 277
among HIV infected and HIV uninfected pregnant women. However, the
numbers of HIV-278 infected women (n=36) compared to HIV-uninfected
participants (n=87) in this study was too 279 small to separately
analyze the diagnostic accuracy of the CES-D by womens HIV status.
280 Further, the study was done at a higher-level health facility,
and pregnant women attending Gulu 281 Regional Referral Hospital
may differ in terms of education, rural vs. urban residence, and
282 access to information from those in the general northern Uganda
population, such that our 283 findings arent more widely
generalizable. 284
In spite of the above stated limitations, our larger PreNAPs
study cohort is designed to 285 have strong internal validity for
comparing differences in psychosocial health variables between 286
HIV infected and uninfected pregnant women. To achieve strong
external validity, future studies 287 could aim to assess the
diagnostic accuracy of the CES-D in the general northern Uganda 288
population, taking a relatively large sample that can permit
separate assessment of this 289 screening tools validity when used
in HIV infected and uninfected pregnant women. In the short 290
term, interventions to mitigate prenatal depression in this
vulnerable pregnant population are 291 needed given the high
prevalence and potential impacts of depression on pregnant women
and 292 their infants. 293
Given the high prevalence of psychiatrist-diagnosed depression
in this population 294 (35.8%) it is worthwhile contemplate
interventions or modifications to current mental health 295
screening policies to mitigate this disorder. Most Ugandan health
facilities at regional referral 296 hospital level have mental
health departments but many opportunities for screening, referral
and 297 follow-up of affected patients are usually missed due to
limited staffing, remuneration, and other 298 incentives to
motivate and retain skilled health workers. It is likely that
screening of pregnant 299 women with the CES-D may help to identify
women who will benefit from treatment for 300
-
Feb 6 13
depression and the CESD is a brief, valid and simple screening
tool to adopt. Although the 301 current national antenatal care
guidelines are silent about integrating screening for mental 302
health problems such as depression into pregnancy care, the high
prevalence of MDDs 303 observed in this study may require revision
of existing ANC guidelines to incorporate screening 304 for
depression at least at the first of, if not all, the four antenatal
care visits. The WHO focused 305 antenatal care guidelines (WHO
2001) recognize the importance of integrating mental health 306
activities into ANC programs and if used together with the WHO
mhGAP (WHO, 2010) for 307 resource poor settings, a number of
primary care mental health interventions targeting pregnant 308
women could be undertaken. 309 310 Conclusions 311
The CES-D scale administered by non-psychiatrists, when compared
to a psychiatrist 312 administered structured MINI interview for
prenatal depression, was found to be a reliable and 313 valid
screening tool in a population of women with and without HIV
attending an antenatal clinic. 314 The CES-D can easily be
administered by nurses and midwives employed by ANC clinics and 315
given the high prevalence (35.8%) of psychiatrist diagnosed
depression in this population, the 316 CES-D could be used to
screen this population for the treatment of depressive disorders.
317 However, further research is needed to assess the time demands
and overall programmatic 318 experience of integrating care for
depression into current ANC service delivery platforms. 319 320
Acknowledgements 321
The authors would like to acknowledge the support of the Gulu
Regional Referral 322 Hospital administration in allowing us to
conduct the study at their hospitals ANC clinic and 323 providing
space for the research team within the hospital buildings. We would
also like to thank 324 the mental health staff at the Mental Health
Unit of Gulu Regional Referral Hospital, especially 325
-
Feb 6 14
Ms. Mary Grace Lanyero, for their willingness to provide
appropriate care for pregnant women 326 diagnosed with current MDD
in this study. We also thank Ms. Sophie Becky Ajok, the 327
coordinator for prevention of mother to child transmission of HIV
(PMTCT) services at GRRH 328 antenatal clinic for helping us
recruit both HIV infected and uninfected pregnant women for this
329 study. The research assistants involved in data collection
including Stella Adoch, Hillary Kilama, 330 Gladys Akello, and
Eunice Asiimwe are also highly appreciated. Lastly, we warmly thank
the 331 mothers for the time they generously gave to the study. 332
333 Funding 334
Funding was provided by the USAID Feed the Future Innovation
Laboratory for 335 Collaborative Research in Nutrition for Africa
(Award Number AID-OAA-L-10-00006 to Tufts 336 University). The
content of this paper is solely the responsibility of the authors
and does not 337 necessarily represent the official views of USAID.
SLY was supported by K01 MH098902 from 338 the National Institute
of Mental Health. The content is solely the responsibility of the
authors and 339 does not necessarily represent the official views
of the National Institute of Mental Health or the 340 National
Institutes of Health. 341 342 Author contributions 343
BKN, SLY, JA, SM, RJS, and JKG conceived and designed study. TOO
conducted MINI 344 Depression diagnostic interviews. BKN and AA
supervised data collection. BKN, JA, SM, JKG, 345 and SLY did the
data analysis. BKN wrote the first draft of the manuscript. All
authors 346 contributed to the interpretation of data, revising the
manuscript critically for important 347 intellectual content; and
final approval of the version to be published. Lastly, we
gratefully 348 acknowledge Patsy Brannon for comments on earlier
drafts. 349 350
-
Feb 6 15
Competing interests 351 The authors declare that they have no
conflict of interest. 352
353 References 354 Akena D, Joska J, Obuku EA, Amos T, Musisi S,
Stein DJ: Comparing the accuracy of brief 355 versus long
depression screening instruments which have been validated in low
and middle 356 income countries: a systematic review. BMC
Psychiatry 2012, 12:187. 357 358 Antelman, G., Kaaya, S., Wei, R.,
Mbwambo, J., Msamanga, G. I., Fawzi, W. W., & Fawzi, M. C. 359
S. (2007). Depressive symptoms increase risk of HIV disease
progression and mortality among 360 women in Tanzania. JAIDS
Journal of Acquired Immune Deficiency Syndromes, 44(4), 470-477.
361 362 Arseniou, S., Arvaniti, A. and Samakouri, M. (2014), HIV
infection and depression. Psychiatry 363 and Clinical
Neurosciences, 68: 96109. doi: 10.1111/pcn.12097 364 365 Bennett,
H. A., Einarson, A., Taddio, A., Koren, G., & Einarson, T. R.
(2004). Prevalence of 366 depression during pregnancy: systematic
review. Obstetrics & Gynecology, 103(4), 698-709. 367 368
Chibanda D, Mangezi W, Tshimanga M, Woelk G, Rusakaniko P,
Stranix-Chibanda L, Midzi S, 369 Maldonado Y, Shetty AK: Validation
of the Edinburgh Postnatal Depression Scale among 370 women in a
high HIV prevalence area in Urban Zimbabwe. Archives of Womens
Mental Heath 371 2010, 13: 201-206. 372 373
-
Feb 6 16
Chishinga N, Kinyanda E, Weiss HA, Patel V, Ayles H, Seedat S:
Validation of brief screening 374 tools for depressive and alcohol
use disorders among TB and HIV patients in primary care in 375
Zambia. BMC Psychiatry 2011, 11(1):75. 376 377 Ciesla JA, Roberts
JE. Meta-analysis of the relationship between HIV infection and the
risk for 378 depressive disorders. Am J. Psychiatry 2001, 158:
725-730 379 380 Coates, J., Swindale, A., & Bilinsky, P.
(2007). Household Food Insecurity Access Scale 381 (HFIAS) for
measurement of food access: indicator guide.Washington, DC: Food
and Nutrition 382 Technical Assistance Project, Academy for
Educational Development. 383 384 Fabiani M, Nattabi B, Pierotti C,
Ciantia F, Opio AA, Musinguzi J, et al. HIV-1 prevalence and 385
factors associated with infection in the conflict-affected region
of North Uganda. Confl Health. 386 2007; 1:3. 387 388 Hatcher,
Abigail M., Alexander C. Tsai, Elias Kumbakumba, Shari L. Dworkin,
Peter W. Hunt, 389 Jeffrey N. Martin, Gina Clark, David R.
Bangsberg, and Sheri D. Weiser. "Sexual Relationship 390 Power and
Depression among HIV-Infected Women in Rural Uganda." PloS one 7,
no. 12 391 (2012): e49821. 392 393 Hwu HG, Joyce PR, Karam EG, Lee
CK, Lellouch J, Wickramaratne PJ, ... & Yeh EK. Cross-394
national epidemiology of major depression and bipolar disorder.
JAMA 1996, 276(0): 293-299 395 396
-
Feb 6 17
Kaharuza FM, Bunnell R, Moss S, Purcell DW, Bikaako-Kajura W,
Wamai N, Downing R, 397 Solberg P, Coutinho A, Mermin J: Depression
and CD4 cell count among persons with HIV 398 infection in Uganda.
AIDS Behav 2006 10 (4 Suppl):S105-111. 399 400 Murray C, Vos T,
Lozano R, et al. Disability-adjusted life years (DALYs) for 291
diseases and 401 injuries in 21 regions, 19902010: a systematic
analysis for the Global Burden of Disease Study 402 2010. Lancet,
380 (2012), pp. 21972223 403 404 Myer L, Smit J, Roux LL, Parker S,
Stein DJ, Seedat S: Common mental disorders among HIV-405 infected
individuals in South Africa; prevalence, predictors, and validation
of brief psychiatric 406 rating scales. AIDS Patient Care STDS
2008, 22(2):147-158. 407 408 Nakasujja, Noeline, Richard L.
Skolasky, Seggane Musisi, Peter Allebeck, Kevin Robertson, 409
Allan Ronald, Elly Katabira, David B. Clifford, and Ned Sacktor.
"Depression symptoms and 410 cognitive function among individuals
with advanced HIV infection initiating HAART in 411 Uganda." BMC
psychiatry 10, no. 1 (2010): 44. 412 413 Nakimuli-Mpungu E, Bass
Jk, Alexandre P, Mills EJ, Musisi S, Ram M, Katabira E, Nachega JB.
414 Depression, Alcohol Use and Adhrerence to Antiretroviral
Therapy in Sub-Saharan Africa. A 415 systematic review. AIDS Behav
2011, 15:376-388. 416 417 Nolen-Hoeksema S. The etiology of gender
differences in depression. In C. Mazure, & G. Keita 418 (Eds.)
(2006), Understanding depression in women. Washington, DC: American
Psychological 419 Association. 420 421
-
Feb 6 18
Olley, B. O., Seedat, S., Nei, D. G., & Stein, D. J. (2004).
Predictors of major depression in 422 recently diagnosed patients
with HIV/AIDS in South Africa. AIDS patient care and STDs, 18(8),
423 481-487. 424 425 Patel V, Araya R, Chowdhary N, King M,
Kirkwood B, Nayak S, Siman G, Weiss HA. Detecting 426 common mental
disorders in primary care in india: a comparison of five screening
427 questionnaires. Psychological Medicine 2008, 38: 221-228. 428
429 Porta, M., Greenland, S., Last, J.M. (Eds.) (2008). A
dictionary of epidemiology. New York, NY: 430 Oxford University
Press. 431 432 Radloff LS: Center for Epidemiologic Studies
Depression Scale (CES-D). In National Institute of 433 Mental
Health, Centre for Epidemiologic Studies. USA: West Publishing
Company; 1977. 434 435 Roberts, B. et al., 2008. Factors associated
with post-traumatic stress disorder and depression 436 amongst
internally displaced persons in northern Uganda. BMC Psychiatry,
8(1), p.38. 437 438 Rochat, T. J., Richter, L. M., Doll, H. A.,
Buthelezi, N. P., Tomkins, A., & Stein, A. (2006). 439
Depression among pregnant rural South African women undergoing HIV
testing. Jama, 295(12), 440 1373-1378. 441 442 Santos, J. R. A.
(1999). Cronbachs alpha: A tool for assessing the reliability of
scales. Journal 443 of extension, 37(2), 1-5. 444 445
-
Feb 6 19
Sheehan DV, Lecrubier Y, Harnett-Sheehan K: The Mini
International Neuropsychiatric Interview 446 (M.I.N.I.): The
Development and Validation of a Structured Diagnostic Psychiatric
Interview. J 447 Clin Psychiatry 1998, 20:2233. 448 449 Simbayi, L.
C., Kalichman, S., Strebel, A., Cloete, A., Henda, N., &
Mqeketo, A. (2007). 450 Internalized stigma, discrimination, and
depression among men and women living with 451 HIV/AIDS in Cape
Town, South Africa. Social science & medicine, 64(9),
1823-1831. 452 453 Strauss ME, Smith GT. Construct Validity:
Advances in Theory and Methodology. Annu. Rev. 454 Clin. Psychol.
NIH Public Access; 2009 Apr;5(1):125 455 456 World Health
Organisation (WHO). 2001. WHO Antenatal Care Randomized Trial:
Manual for 457 the Implementation of the New Model. Geneva. WHO.
WHO/RHR/01.30. 458 459 World Health Organisation (WHO): mhGAP
intervention guide for mental, neurological and 460 substance use
disordered in non-specialised health settings. Geneva, Switzerland.
WHO Press 461 2010 462 463 Wingood GM, & DiClemente RJ.
Application of the theory of gender and power to examine HIV-464
related exposures, risk factors, and effective interventions for
women. Health education & 465 behavior, 2000, 27(5): 539-565.
466 467 468 469
-
Feb 6 20
Table 1. Characteristics of 123 pregnant participants in the
study to validate the Center for 470 Epidemiologic
Studies-Depression (CES-D) scale 471 Variable Parameter/category
Total (N=123) Socio-demographics Participants age, years Median
(IQR) 23 (20-27) Marital status, (%) Single 6.5
Married/cohabiting 87.8 Widowed 2.4 Separated 3.3
Education, (%) No education 4.1 Primary 49.6 O level only 31.7 A
levels or higher 14.6
Occupation, (%) Housewife 50.4 Farmer 8.9 Trader 22.8 Formally
employed 13.0 Other 4.9
Language, (%) Acholi 90.2 Langi 4.1 Other 5.
Gestational age at interview, weeks Median (IQR) 19 (15-21)
HIV-infected, (%) Positive 29.3 Monthly household income per capita
(in Uganda Shillings (UGX) and US dollars (USD)
Median (IQR) UGX, 50,000 (25,000-120,000); USD, 20 (10-48)
IFIAS score Median (IQR) 9 (5-13) IFIAS category, (%) Food
secure 3.3
Mild food insecure 20.3 Moderate food insecure 24.4 Severe food
insecure 52.0
CES-D Score Mean (SD) 17.1(13.8) Range (Min-Max) 55 (0-55)
MINI-defined current major depressive disorder (%)
Yes 35.8
472 473 474 475
-
Feb 6 21
Table 2. Unadjusted bivariate associations between CES-D scores
and selected variables 476 (N=123) 477 Factor Category Unadjusted
coefficient P value
HIV status HIV negative Reference
HIV positive 8.64272 0.001
IFIAS score N/A 1.029777 0.000
Per capita monthly income First tertile Reference
Second tertile -5.585366 0.066
Third tertile -6.644512 0.030
478 479 480 481 482 483 484 485 486 487 488 489 490 491 492
-
Feb 6 22
Figure 1. Median CES-D Scores by Psychiatric Diagnosis using the
MINI 493
CES-
D Sc
ore
05
1015
20p 5
0 of c
esds
core
No Yes
Current MDD diagnosis by MINI depression derived outputs
494 495 496 497 498 499 500 501 502 503 504
-
Feb 6 23
Figure 2. Receiver Operating Characteristic Curve for the CES-D
total scores for diagnosis of 505 current MDD 506
507 508 509 510 511 512 513 514 515 516 517 518 519 520
521 522 523 524 525 526 527 528
0.00
0.00
0.000.25
0.25
0.250.50
0.50
0.500.75
0.75
0.751.00
1.00
1.00Sensitivity
Sens
itivity
Sensitivity0.00
0.00
0.000.25
0.25
0.250.50
0.50
0.500.75
0.75
0.751.00
1.00
1.001 - Specificity
1 - Specificity
1 - SpecificityArea under ROC curve = 0.8277
Area under ROC curve = 0.8277
Area under ROC curve = 0.8277
-
Feb 6 24
Figure 3. Sensitivities and positive predictive values for the
CES-D by cut-off scores for 529 diagnosis of current MDD 530
531
Se/P
PV
TOTAL CES-D SCORES
532