Using the Beck Anxiety Inventory Among South Africans Living With HIV

Abstract

We administered the Beck Anxiety Inventory (BAI) to 101 adults receiving HIV treatment. Exploratory factor analysis yielded a single anxiety factor that accounted for 68.7% of the variance in the data. A single score may be used to indicate the overall level of anxiety of individuals receiving HIV treatment in South Africa.

Keywords

Beck Anxiety Inventory HIV South Africa exploratory factor analysis

The ubiquity of common mental disorders worldwide has become increasingly salient as numerous epidemiological studies have yielded results indicating high prevalence rates for these conditions (World Health Organization [WHO], 2008). In low- and middle-income countries (LMICs) as well, an increasing number of studies have shown high rates of mental disorders, for example, depression, anxiety, posttraumatic stress, and substance abuse (Bradshaw et al., 2003; Desajarlais, Eisenberg, Good, & Kleinman, 1995; Herman et al., 2009; Lopez, Mathers, Ezzati, Jamison, & Murray, 2006; WHO, 2001). While depression has received special scrutiny in view of its association with numerous other problems such as worker absenteeism, suicidal behavior, poor physical health, poor quality of life, and suboptimal child care and family functioning (Gotlib & Hammen, 2009), anxiety as a psychological phenomenon has received relatively little attention. Yet anxiety negatively affects quality of life, health outcomes, and personal well-being (Antony & Stein, 2009).

Anxiety Among Persons Living With HIV

Anxiety among persons living with HIV has been fairly well documented with a variety of assessment instruments. For example, using the Spielberger State-Trait Anxiety Scale, Coleman and Holzemer (1999) found moderate levels of anxiety among a sample of African American men and women living with HIV. Among outpatients receiving care at a Quebec hospital, 43.2% were considered potential cases and 21.6% were likely cases of anxiety according to the Hospital Anxiety and Depression Scale, indicated by scores of 8 or higher and 10 or higher, respectively (Savard, Laberge, Gauthier, Ivers, & Bergeron, 1998). Similarly, among a sample of 98 HIV-positive persons in Vermont, New Hampshire, Massachusetts, and Maine, the mean score on the Beck Anxiety Inventory (BAI; Beck, Epstein, Brown, & Steer, 1988) was 16.21 (SD = 13.05), indicating on average a mild to moderate level of anxiety (Gonzalez, Solomon, Zvolensky, & Miller, 2009). In a two-group intervention study, the BAI means of a sample of persons living with HIV were 13.08 (SD = 9.95) and 13.22 (SD = 10.19) at baseline and 9.59 (SD = 9.43) and 10.32 (SD = 9.57) after cognitive therapy and acceptance and commitment therapy, respectively (Forman, Herbert, Moitra, Yeomans, & Geller, 2007). These findings indicate nontrivial levels of anxiety and suggest that the BAI is able to detect change over time.

In the 1990s, antiretroviral treatment (ART) was not as widely available as it has become in recent years (Mbali, 2013). As a result, over the past two decades HIV has changed from being a life-threatening condition to a manageable chronic illness. This transition may account for lower rates of anxiety symptoms among infected persons.

Anxiety among persons living with HIV in LMICs is under-researched. Among the screening instruments that have been used in sub-Saharan Africa are the Edinburgh Postnatal Depression Scale (Rochat et al., 2006), the Hopkins Symptom Checklist (Kagee, 2010), and the Center for Epidemiologic Studies Depression Scale (Myer et al., 2008). However, to our knowledge, other than research by Nel and Kagee (2011) and Steele and Edwards (2008), the BAI has almost never been used in African settings. Nel and Kagee (2011) found high internal consistency of the BAI among ART users in the Western Cape, and Steele and Edwards (2008) found high internal consistency of the Xhosa version of the BAI.

The Beck Anxiety Inventory

The BAI is a self-report inventory containing 21 multiple-choice items in which the respondent is asked to rate the severity of various symptoms pertaining to anxiety in terms of how he or she has been feeling in the last week (Beck et al., 1988). The scale is designed for persons 17 years and older. Each symptom is presented with four possible response options, and the respondent is asked to endorse the one that applies to him or her best. The response options are not at all, mildly, moderately, and severely. The minimum score is 0 and the maximum obtainable score is 63 (Beck et al., 1988). A score above 10 on the BAI is considered to indicate mild anxiety, and a score above 19 indicates moderate anxiety (Hoyer, Becker, Neumar, Soeder, & Margraf, 2000).

The BAI was developed in the 1980s in response to a need for an instrument that could discriminate between symptoms of anxiety and depression but simultaneously retain convergent validity (Beck et al., 1988). The original validation study was conducted using three samples of American psychiatric outpatients (n = 1,086) who were recruited between 1980 and 1986. The validation of the BAI was conducted among a sub-sample who had been diagnosed with major depressive disorder, dysthymic disorder, atypical depression, panic disorder, generalized anxiety disorder, agoraphobia, social and simple phobia, and various other problems such as academic difficulties and adjustment disorders (Beck et al., 1988).

The BAI has come to be known as a valid and reliable measure of anxiety with high internal consistency, test–retest reliability and good concurrent and discriminant validity. It is a widely used measure of anxiety in behavioral research and practice and is often used to screen for anxiety disorders in health care settings. It has been widely applied in various cultural settings in the United States and elsewhere in the world (e.g., Beck, 1991; Creamer, Foran, & Bell, 1995; Fydrich, 1992; Kabacoff, Segal, Hersen, & Van Hasselt, 1997; Novy, Stanley, Averill, & Daza, 2001; Osman, Kopper, Barrios, Osman, & Wade, 1997; Saemundsson et al., 2011; Steer, & Ranieri, 1993). Its psychometric properties have been shown to be uniformly robust in a variety of geographical contexts.

In South Africa, a Xhosa version of the BAI was developed by Steele and Edwards (2008), the psychometric properties of which were found to be acceptable. In this study, Cronbach’s alpha of the instrument was .92, indicating high internal consistency and comparing favorably with the original BAI validation studies (Steele & Edwards, 2008). The item-to-total correlations of the Xhosa version ranged between .44 and .71, providing evidence of the coherence of the BAI as a measure of a single construct.

Factor Structure of the BAI

In an exploratory factor analysis (EFA) of the BAI, Hewitt and Norton (1993) found two factors that corresponded to cognitive and somatic symptoms. In another study using principal-factor analysis and oblique rotation of the combined items of the Beck Depression Inventory and the BAI, Steer, Beck, Clark, and Ranieri (1995) identified two correlated factors, namely, Depression and Anxiety. After a Schmid–Leiman transformation of these data, a single higher order factor approximating distress or negative affect explained twice the amount of variance than the two factors together. In a study of the BAI among older medical patients, confirmatory factor analysis showed that a second-order four-factor structure was the best model fit (Wetherell & Arean, 1997). The authors concluded that although disadvantaged older persons experience similar dimensions of anxiety as younger persons, the constellations of symptoms may differ. For example, fear of dying may be experienced to a greater extent among older persons compared to their younger counterparts. In the present study, we sought to determine the performance of items and identify the factor structure of the BAI among a sample of persons living with HIV receiving ART at a public health clinic in South Africa.

Method

Participants

Participants were patients receiving ART at an infectious diseases clinic of a major public hospital in the Western Cape province of South Africa. The hospital, located approximately 30 km east of Cape Town, is a tertiary institution and is operated under the auspices of the Department of Health of the Western Cape Provincial Government. Participants were recruited by means of convenience sampling. The only criterion for exclusion was diagnosis with a serious mental illness or developmental delay. Among the sample, no participants were excluded on this basis.

Procedures

As patients arrived for their medical appointments, clinic nurses informed them of the study, handed them a pamphlet describing the study, and informed them that if they wished to learn more about the research, they could approach members of the research team stationed in a private office at the clinic. When participants indicated an interest in meeting one of the researchers, the nurse escorted the patient to the researcher’s office. On meeting each patient, the researcher explained the study and invited patients to participate. Those who agreed were asked to provide informed consent and complete the questionnaire battery, which included the BAI. Participants who had difficulty completing the questionnaires, for example, those with low levels of literacy, were encouraged to seek assistance from the researchers, who then verbally administered the questionnaires. Only a small number of participants (less than 10% of the sample) requested assistance with one or two questionnaire items in this regard. Those who indicated that they had been diagnosed with bipolar disorder, schizophrenia, or related psychotic disorders were excluded from the study. Participants who completed the questionnaire were given a grocery voucher to the value of ZAR50 (US$5) as a token of appreciation for their time and effort. Ethics approval for the study was obtained from the Stellenbosch University Health Research Ethics Committee. No patients who were invited to participate declined.

Measures

Demographic Information

A demographic questionnaire asked participants to provide information on age, gender, marital status, first language, and employment status.

Anxiety

The BAI, a 21-item self-report scale, was used to measure severity of symptoms of anxiety in the sample. Beck and Steer (1993) validated the scale among samples within the United States and found high internal consistency, with alpha coefficients ranging from .92 to .94. The test demonstrated high content, concurrent, construct, discriminant, and factorial validities for populations within the United States. As mentioned earlier, Steele and Edwards (2008) also investigated the psychometric properties of a Xhosa translation of the BAI in South Africa and found that the internal consistency, item–total correlations, means, and standard deviations were comparable to the validation studies conducted in the United States.

Translation of the BAI

The development and validation of the Xhosa version of the BAI has been described by Steele and Edwards (2008). For the purpose of the present study, the BAI was translated into Afrikaans by a bilingual member of the research team (AN). The translated scale was sent to the Stellenbosch University Language Centre for back-translation and authentication. Feedback from the authentication process indicated that the Afrikaans version differed negligibly from the original English scales in terms of content and meaning. Following these procedures, the BAI was made available to participants in English, Afrikaans, and Xhosa. The majority of participants responded to the questionnaire in English.

Statistical Analysis

The data were entered into SPSS version 20.0. We used EFA to analyze the data and to determine the underlying factor structure of the data. The method of extraction was maximum likelihood factor analysis based on eigenvalues greater than 1 (Kaiser, 1960) as well as the scree plot. A total of three factors emerged from the data. The factors were rotated with the Promax procedure (κ = 4) to an oblique solution. The correlations between the factors that were obtained from the first-order analysis formed part of a higher order factor analysis. This higher order factor solution was then transformed to an orthogonal Schmid–Leiman hierarchical factor solution (Schmid & Leiman, 1957). We obtained the SPSS syntax to calculate the Schmid–Leiman hierarchical factor solution from Wolff and Preising (2005).

Results

Description of the Sample

The sample consisted of 101 participants of which 18 (18%) were male and 83 (82%) female. No patient who was approached to participate declined. The mean age of the sample was 35 years (SD = 7.1). The participants were almost exclusively Black and “Colored,”¹ with 55.4% and 31.7% indicating Xhosa and Afrikaans as their first language, respectively, and English as their second language. A large minority of participants (43%) indicated they were unemployed. Table 1 presents demographic information of the sample.

Table 1.

Demographic Characteristics of the Sample.

Variable	No. of Respondents	% of Total Sample
Gender
Male	18	17.8
Female	83	82.2
Age (years)
20–29	20	29.7
30–39	44	43.6
40–49	24	23.8
50–51	3	3.0
First language
Xhosa	56	55.4
Afrikaans	32	31.7
Other	13	12.9
Current work situation
Employed full-time	29	28.7
Employed part-time	17	16.8
Unemployed	43	42.6
Homemaker	4	4.0
Student	4	4.0
Disabled	4	4.0
Annual family income, ZAR
<10,000	40	41.7
10,001–40,000	16	16.7
40,001–80,000	3	3.1
80,001–110,000	1	1.0
110,001–170,000	0	0.0
170,001–240,000	1	1.0
≥240,000	1	1.0
Do not know	34	35.4

Note. ZAR10 = US$1.

EFA of the BAI

On the first iteration of the EFA, inspection of the correlation matrix (R matrix) yielded low correlations (<.8) among the items, indicating no redundancy. The determinant was 4.093 × 10⁻⁵, further indicating that multicollinearity was not a problem (Field, 2005). The Kaiser–Meyer–Olkin measure of sampling adequacy was .76, a value close to 1.00, indicating that factor analysis was a suitable method of analysis (Field, 2005). Bartlett’s test of sphericity was significant (p < .01) suggesting that the correlation matrix was significantly different from the identity matrix. This finding indicated that the items correlated with one another, making the possibility of finding factors more likely (Field, 2005).

Primary Factor Structure

After the EFA was conducted, a total of three primary factors (based on the scree plot and eigenvalues >1) were extracted, accounting for 41.24% of the variance among the items. The eigenvalues of the factors were 6.95, 1.70, and 1.52, respectively. We then conducted a higher order factor analysis with a single second-order factor. The higher order factor solution was then transformed to an orthogonal Schmid–Leiman hierarchical factor solution (Schmid & Leiman, 1957), which produced a single second-order factor and three uncorrelated group or primary factors. The Schmid–Leiman solution (SLS) allows for a clear evaluation of the relative influences of factors at different levels of the factor hierarchy (McDonald, 1999). As can be seen from Table 2, the SLS produced a single well-defined second-order or general factor, which we named Anxiety. It should be noted that one item did not load on any of the three factors after the SLS was conducted. Thus, although three factors were identified, the results show that 68.7% of the variance was explained by the secondary factor alone. The second-order factor accounts for a greater proportion of the variance than each of the other lower level factors extracted. On the basis of these results, it may be concluded that the BAI may be treated as a unidimensional scale among this sample of ART users, thus validating the calculation of a total score with a minimum or 0 and a maximum of 63.

Table 2.

Factor Loadings of Schmid–Leiman Solution for the BAI (Cutoff = .30).

Item	General Factor	Factor 1	Factor 2	Factor 3
BAI1	.256	.267	−.169	.132
BAI2	.478	.228	.048	.239
BAI3	.478	.269	−.038	.286
BAI4	.454	−.066	−.013	.888
BAI5	.445	.101	.335	.004
BAI6	.594	.401	.137	−.06
BAI7	.419	.255	.11	−.006
BAI8	.502	.294	.158	−.022
BAI9	.374	.198	.199	−.1
BAI10	.391	.151	.238	−.035
BAI11	.397	.117	.106	.244
BAI12	.508	.43	.005	−.056
BAI13	.633	.487	−.012	.056
BAI14	.516	.172	.292	.039
BAI15	.45	.357	.03	−.041
BAI16	.387	−.06	.521	−.048
BAI17	.542	−.148	.68	.14
BAI18	.441	.109	.132	.291
BAI19	.599	.497	−.061	.058
BAI20	.514	.238	.252	−.035
BAI21	.378	.249	−.052	.188
% variance explained	68.7	13.8	8.9	8.6

Note. BAI = Beck Anxiety Inventory.

Limitations of the SLS

In their article discussing the relationship between higher order factor models and hierarchical factor models, Yung, Thissen, and McLeod (1999) note that the SLS often results in a restricted hierarchical model and that structural equation modeling may yield a less rigid hierarchical factor solution. Yet we found that the SLS was a relatively simple and useful method to use to gain a further perspective on the relations between variables and higher order factors (Wolff & Preising, 2005) and that further analysis of our results would not necessarily change the overall outcome.

Internal Consistency of the BAI and Endorsement of Items

The BAI showed high internal consistency among the sample (α = .89). The mean score on the BAI was 11.60 (SD = 10.15), indicating that on average the sample had mild symptoms of anxiety. In terms of severity of anxiety, 50.6% of the sample fell in the minimal range, 24.1% in the mild range, 19.5% in the moderate range, and 5.7% fell in the severe range.

Table 3 shows the item-by-item descriptive analysis of the scale and includes the percentage of respondents who endorsed each response option for each item. In the far right column, the percentage of the sample endorsing the “moderately” and “severely” response options were collapsed into one column to give an indication of which items were reported to a greater extent by the population as a whole. More than a quarter of the population endorsed the items “Fear of the worst happening” (38.4%) and “Fear of dying” (26.8%) as affecting them moderately or severely. These items were clearly related to the salience of mortality they experienced, presumably related to their HIV status. More than one fifth of the sample endorsed the following items: “Numbness or tingling,” “Wobbliness in legs,” “Nervous,” and “Indigestion or discomfort in abdomen.” These items were all related to bodily sensations. Other items that were endorsed by almost 20% of the sample were “Feeling hot,” “Dizzy and lightheaded,” “Heart pounding or racing,” “Fear of losing control,” and “Scared.” The items that were endorsed by the least number of participants as affecting them moderately or severely were “Unsteady” (8.1%), “Faint” (8.0%), “Face flushed” (10.9%), “Shaky” (11.1%), and “Feelings of choking” (12.0%).

Table 3.

Item Analysis of BAI Items.

Item	Description	M	SD	Not at All	Mildly	Moderately	Severely	Moderately + Severely
1	Numbness or tingling	0.69	0.85	54	25	19	2	21
2	Feeling hot	0.69	0.87	53.5	27.7	14.9	4	18.9
3	Wobbliness in legs	0.63	0.93	62.6	16.2	16.2	5.1	21.3
4	Unable to relax	0.59	0.84	59	28	8	5	13
5	Fear of the worst happening	1.1	1.16	45.5	16.2	21.2	17.2	38.4
6	Dizzy or lightheaded	0.63	0.86	59.4	20.8	16.8	3	19.8
7	Heart pounding or racing	0.66	0.93	59.6	21.2	13.1	6.1	19.2
8	Unsteady	0.39	0.7	70.7	21.2	6.1	2	8.1
9	Terrified	0.58	0.87	62.6	22.2	10.1	5.1	15.2
10	Nervous	0.79	1.02	53.1	26.5	9.2	11.2	20.4
11	Feelings of choking	0.4	0.75	74	14	10	2	12
12	Hands trembling	0.55	0.89	65.3	19.8	8.9	5.9	14.8
13	Shaky	0.37	0.76	76.8	12.1	8.1	3	11.1
14	Fear of losing control	0.61	1.07	70.3	10.9	5.9	12.9	18.8
15	Difficulty breathing	0.44	0.85	74.3	11.9	8.9	5	13.9
16	Fear of dying	0.81	1.12	59.4	13.9	12.9	13.9	26.8
17	Scared	0.62	0.91	60.6	22.2	11.1	6.1	17.2
18	Indigestion or discomfort in abdomen	0.79	0.95	51.5	23.8	18.8	5.9	24.7
19	Faint	0.28	0.73	83.2	8.9	4	4	8
20	Face flushed	0.38	0.79	76.2	12.9	6.9	4	10.9
21	Sweating (Not due to heat)	0.54	0.91	67.3	17.8	7.9	6.9	14.8

Note. BAI = Beck Anxiety Inventory.

Discussion

The sample was predominantly Black and Colored South Africans, many of whom were unemployed and living under low socioeconomic conditions. The primary factor structure consisted of three factors, although most of the eigenvalues were only trivially greater than 1.0, indicating limited stability of the factors. For this reason, it was appropriate to conduct a hierarchical factor solution following a higher order factor analysis that yielded a single general factor accounting for 68.7% of the variance. As a result of the higher order factor analysis, it appears that the calculation of a single anxiety score is valid among the population under study. These findings are in contrast with other studies that found a two-factor solution, namely, cognitive and somatic (Hewitt & Norton, 1993). Yet, in the study by Hewitt and Norton (1993), the two factors were highly correlated with one another. The identification of a single higher order anxiety factor among our sample appears unique in the existing factor analytic literature on the BAI.

Of considerable interest was the sample’s endorsement of the scale’s individual items. As can be expected of a population living with a serious and potentially fatal illness, the items endorsed by the highest number of participants were “Fear of the worst happening” and “Fear of dying.” Despite the success of antiretroviral drugs in maintaining the health of persons living with HIV, it is likely that many persons may be concerned about their mortality, which for them is dying of AIDS. The somatic items were also endorsed among many members of the sample, although not as highly as the items involving fear. The scores of half of the sample fell in categories of the BAI that were above the minimal range, indicating that anxiety for many persons living with HIV may be a psychological experience that features prominently in their lives. To this extent, the imperative to make psychological services available to ameliorate distress, anxiety, and other common mental health problems needs to be considered in conjunction with the limitations on resources, skills, and personnel required for such services in public health systems in LMICs.

The study was conducted in an infectious diseases clinic of a hospital outside of Cape Town in South Africa among a sample of 101 participants. Thus, the generalizability of the data beyond the population from which the sample was recruited is limited. Furthermore, while the BAI was made available to respondents in English, Afrikaans, and Xhosa, some participants who indicated difficulty in reading some of the test items were assisted by a research assistant. Although negligible, social desirability bias may have accounted for the test performance among these participants.

Our data are not directly comparable with those of Steele and Edwards (2008) as these authors did not seek to determine the factor structure of the BAI nor did they report on the performance of individual items beyond the means, standard deviations, and corrected item–total correlations. Nonetheless, the mean score of our sample (11.60, SD = 10.15) was nonsignificantly different from that of Steele and Edwards (2008; 14.94, SD = 11.61), even though their sample consisted of individuals receiving psychological or psychiatric care.

There are two important implications of the study for clinical practice. The data indicate that the BAI may be used as a unidimensional scale, thus permitting the calculation of a total score with a minimum or 0 and a maximum of 63. To this extent, the BAI may be used as a way to identify patients in need of support and assistance in ameliorating symptoms of anxiety. Kagee (2010) has shown that a small but nontrivial proportion of persons living with HIV may experience elevated anxiety. The data from the present study indicate that the BAI may be appropriately used to identify such persons with a view to making treatment available in contexts where resources are available.

A second implication of the study pertains to the items that were most commonly endorsed by the sample, that is, items related to the salience of mortality in the context of living with HIV. In the context of increasing availability of ART, living with HIV has come to be regarded as a chronic rather than a terminal medical condition. Yet stigma often acts as a barrier to appropriate disclosure and medication adherence (Coetzee, Kagee, & Vermeulen, 2011), which in turn may constitute a threat to survival rates. Nonetheless, the BAI may assist in identifying persons experiencing mortality-related symptoms of anxiety with a view to their amelioration and enhancement of adherence to their medication regimens.

The present study also has implications for future research. The debate regarding the utility of routine screening for common mental disorders of medical patients is unsettled, and cautions have been raised concerning the sensitivity and specificity self-report measures when compared to the gold standard of a structured clinical interview (see Kagee, Tsai, Lund, & Tomlinson, 2013, for a review). A yet-unanswered empirical question is whether elevated scores on the BAI accurately predict caseness for generalized anxiety disorder among persons living with HIV as determined by instruments such as the Structured Clinical Interview for the Diagnostic and Statistical Manual for Mental Disorders (First, Spitzer, Gibbon, & Williams, 2002) and the Composite International Diagnostic Interview (Kessler & Ustün, 2004). Studies on the sensitivity and specificity of the BAI in comparison with gold standard instruments among persons living with HIV in LMICs await investigation.

In the context of a limited number of self-report measures to assess anxiety among persons living with HIV, we believe the present study adds to the literature on the BAI among this population. The study brings into focus the need for assessment of the psychological experiences of persons living with HIV as it is closely related to their well-being and quality of life.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes

Author Biographies

Ashraf Kagee is Professor of Psychology at Stellenbosch University.

Bronwyne Coetzee is a doctoral student in the Department of Psychology at Stellenbosch University.

Wylene Saal is a doctoral student in the Department of Psychology at Stellenbosch University.

Adriaan Nel received his masters degree in Psychology at Stellenbosch University.

References

Antony

M. M.

Stein

M. B.

(Eds.). (2009). Oxford handbook of anxiety and related disorders. New York, NY: Oxford University Press.

Beck

A. T.

(1991). Relationship between the Beck Anxiety Inventory and the Hamilton Anxiety Rating Scale with anxious outpatients. Journal of Anxiety Disorders, 5, 213–223.

Beck

A. T.

Epstein

Brown

Steer

R. A.

(1988). An inventory for measuring clinical anxiety: Psychometric properties. Journal of Consulting and Clinical Psychology, 56, 893–897. doi:10.1037/0022-006X.56.6.893

Beck

A. T.

Steer

R. A.

(1993). Beck Anxiety Inventory Manual. San Antonio, TX: Harcourt Brace.

Bradshaw

Groenewald

Laubscher

Nannan

Nojilana

Norman

. . . Schneider

(2003). Initial burden of disease estimates for South Africa, 2000. Cape Town, South Africa: Medical Research Council.

Coetzee

Kagee

Vermeulen

(2011). Structural barriers to adherence to antiretroviral therapy in a resource-constrained setting: The perspectives of health care providers. AIDS Care, 23, 146–151. doi:10.1080/09540121.2010.498874

Coleman

C. L.

Holzemer

W. L.

(1999). Spirituality, psychological well-being, and HIV symptoms for African Americans living with HIV disease. Journal of the Association of Nurses in AIDS Care, 10(1), 42–50.

Creamer

Foran

Bell

(1995). The Beck Anxiety Inventory in a non-clinical sample. Behaviour Research and Therapy, 33, 477–485. doi:10.1016/0005-7967(94)00082

Desajarlais

Eisenberg

Good

Kleinman

(1995). World mental health: Problems and priorities in low income countries. New York, NY: Oxford University press.

10.

Field

A. P.

(2005). Discovering statistics using SPSS: (and sex, drugs and rock ‘n’ roll). London, England: Sage.

11.

First

M. B.

Spitzer

R. L.

Gibbon

Williams

J. B. W.

(2002). Structured clinical interview for DSM-IV-TR Axis I disorders, research version, patient edition (SCID-I/P). New York: Biometrics Research, New York State Psychiatric Institute.

12.

Forman

E. M.

Herbert

J. D.

Moitra

Yeomans

P. D.

Geller

P. A.

(2007). A randomized controlled effectiveness trial of acceptance and commitment therapy and cognitive therapy for anxiety and depression. Behavior Modification, 31, 772–799. doi:10.1177/0145445507302202

13.

Fydrich

(1992). Reliability and validity of the Beck Anxiety Inventory. Journal of Anxiety Disorders, 6, 55–61. doi:10.1016/0887-6185(92)90026-4

14.

Gonzalez

Solomon

S. E.

Zvolensky

M. J.

Miller

C. T.

(2009). The interaction of mindful-based attention and awareness and disengagement coping with HIV/AIDS-related stigma in regard to concurrent anxiety and depressive symptoms among adults with HIV/AIDS. Journal of Health Psychology, 14, 403–413. doi:10.1177/1359105309102193

15.

Gotlib

I. H.

Hammen

C. L.

(Eds.). (2009). Handbook of depression (2nd ed.). New York, NY: Guilford Press.

16.

Herman

A. A.

Stein

D. J.

Seedat

Heeringa

S. G.

Moomal

Williams

D. R.

(2009). The South African stress and health (SASH) study: 12-month and lifetime prevalence of common mental disorders. South African Medical Journal, 99, 339–344.

17.

Hewitt

P. L.

Norton

G. R.

(1993). The Beck Anxiety Inventory: A psychometric analysis. Psychological Assessment, 5, 408–412. doi:10.1037/1040-3590.5.4.408

18.

Hoyer

Becker

E. S.

Neumar

Soeder

Margraf

(2000). Screening for anxiety in an epidemiological sample: Predicting accuracy of questionnaires. Anxiety Disorders, 16, 113–134.

19.

Kabacoff

R. I.

Segal

D. L.

Hersen

Van Hasselt

V. B.

(1997). Psychometric properties and diagnostic utility of the Beck Anxiety Inventory and the state-trait anxiety inventory with older adult psychiatric outpatients. Journal of Anxiety Disorders, 11, 33–47. doi:10.1016/S0887-6185(96)00033-3

20.

Kagee

(2010). Psychological distress among persons living with HIV, hypertension, and diabetes. AIDS Care, 22, 1517–1521. doi:10.1080/09540121.2010.484458

21.

Kagee

Tsai

Lund

Tomlinson

(2013). Screening for common mental disorders in low resource settings: Reasons for caution and a way forward. International Health, 5, 11–14. doi:10.1093/inthealth/ihs004

22.

Kaiser

H. F.

(1960). The application of electronic computers to factor analysis. Educational and Psychological Measurement, 20, 141–151. doi:10.1177/001316446002000116

23.

Kessler

R. C.

Ustün

T. B.

(2004). The World Mental Health (WMH) Survey Initiative version of the World Health Organization (WHO) Composite International Diagnostic Interview (CIDI). International Journal of Methods in Psychiatric Research, 13, 93–121.

24.

Lopez

A. D.

Mathers

C. D.

Ezzati

Jamison

D. T.

Murray

C. J. L.

(2006). Global burden of disease and risk factors. New York, NY: Oxford University Press and the World Bank.

25.

Mbali

(2013). South African aids activism and global health politics. New York, NY: Palgrave MacMillan.

26.

McDonald

R. P.

(1999). Test theory: A unified treatment. Mahwah, NJ: Lea.

27.

Myer

Smit

Le Roux

Parker

Stein

D. J.

Seedat

(2008). Common mental disorders among HIV-infected individuals in South Africa: Prevalence, predictors, and validation of brief psychiatric rating scales. AIDS Patient Care and STDs, 22, 147–158. doi:10.1089/apc.2007.0102

28.

Nel

Kagee

(2011). Common mental health problems and antiretroviral therapy adherence. AIDS Care, 23, 1360–1365. doi:10.1080/09540121.2011.565025

29.

Novy

D. M.

Stanley

M. A.

Averill

Daza

(2001). Psychometric comparability of English- and Spanish-language measures of anxiety and related affective symptoms. Psychological Assessment, 13, 347–355. doi:10.1037/1040-3590.13.3.347

30.

Osman

Kopper

B. A.

Barrios

F. X.

Osman

J. R.

Wade

(1997). The Beck Anxiety Inventory: Reexamination of factor structure and psychometric properties. Journal of Clinical Psychology, 53, 7–14. doi:10.1002/(SICI)1097-4679(199701)53:1<7: AID-JCLP2>3.0.CO;2-S

31.

Rochat

T. J.

Richter

L. M.

Doll

H. A.

Buthelezi

N. P.

Tomkins

Stein

(2006). Depression among pregnant rural South African women undergoing HIV testing. Journal of the American Medical Association, 295, 1373–1378. doi:10.1001/jama.295.12.1376

32.

Saemundsson

B. R.

Porsdottir

Kristjansdottir

Olason

D. P.

Smari

Sigurosson

J. F.

(2011). Psychometric properties of the Icelandic Version of the Beck Anxiety Inventory in a clinical and a student population. European Journal of Psychological Assessment, 27, 133–141. doi:10.1027/1015-5759/a000059

33.

Savard

Laberge

Gauthier

J. G.

Ivers

Bergeron

M. G.

(1998). Evaluating anxiety and depression in HIV-infected patients. Journal of Personality Assessment, 71, 349–367.

34.

Schmid

Leiman

J. M.

(1957). The development of hierarchical factor solutions. Psychometrika, 22, 53–61. doi:10.1007/BF02289209

35.

Steele

G. I.

Edwards

D. J. A.

(2008). Development and validation of the Xhosa translations of the Beck inventories: 2. Item analysis, internal consistency and factor analysis. Journal of Psychology in Africa, 18, 207–216.

36.

Steer

R. A.

Beck

A. T.

Clark

D. A.

Ranieri

W. F.

(1995). Common and specific dimensions of self-reported anxiety and depression: A replication. Journal of Abnormal Psychology, 104, 542–545. doi:10.1037//0021-843X.104.3.542

37.

Steer

R. A.

Ranieri

W. F.

(1993). Further evidence for the validity of the Beck Anxiety Inventory with psychiatric outpatients. Journal of Anxiety Disorders, 7, 195–205. doi:10.1016/0887-6185(93)90002-3

38.

Wetherell

J. L.

Arean

P. A.

(1997). Psychometric evaluation of the Beck Anxiety Inventory with older medical patients. Psychological Assessment, 9, 136–144. doi:10.1037/1040-3590.9.2.136

39.

Wolff

H. G.

Preising

(2005). Exploring item and higher order factor structure with the Schmid-Leiman solution: Syntax codes for SPSS and SAS. Behavior Research Methods, 37, 48–58. doi:10.3758/BF03206397

40.

World Health Organization. (2001). World health report, 2001, mental health: New understanding, new hope. Geneva, Switzerland: Author.

41.

World Health Organization. (2008). Cross-national comparisons of the prevalences and correlates of mental disorders: WHO international consortium in psychiatric epidemiology. Bulletin of the World Health Organization, 78, 413–426.

42.

Yung

Y. F.

Thissen

McLeod

L. D.

(1999). On the relationship between the higher-order factor model and the hierarchical factor model. Psychometrika, 64, 113–128. doi:10.1007/BF02294531