Development and Evaluation of a Complementary and Alternative Medicine Use Survey in African-Americans with Acquired Immune Deficiency Syndrome

Abstract

Objectives:

The purpose of the current study was to develop and evaluate the psychometric properties of a culturally- and stage-of-disease-appropriate measure of complementary and alternative medicine (CAM) use among a population of African-American individuals with acquired immune deficiency syndrome (AIDS) using a mixed-method design.

Design:

Data were collected in two phases. In phase 1, qualitative data were used to refine an existing CAM measure for the specific study population in the present study. In phase 2, this refined instrument was implemented in a larger sample. The resulting numeric data were analyzed to evaluate the psychometric properties of the revised CAM instrument.

Setting:

Data were collected from patients who were receiving care from the infectious disease clinic of a large, public, urban hospital in the Southeastern United States.

Subjects:

Patients were eligible to participate if they (1) were receiving their care from the clinic, (2) had an AIDS diagnosis, (3) were identified as African-American, (4) were ≥21 years of age, (5) spoke English, and (6) were not cognitively impaired.

Measures:

Focus groups in phase 1 were conducted with a semistructured focus group guide. Participants also completed a basic sociodemographic survey. Phase 2 participants used programmed laptops to answer questions about their CAM use and several sociodemographic questions.

Results:

Information from the focus groups prompted some substantive revisions in the already-existing CAM survey. The revised instrument had satisfactory face validity and adequate test–retest reliability (r = 0.79). Furthermore, the instrument factored in a manner that was interpretable and consistent with prior findings.

Conclusions:

In order for human immunodeficiency virus health care providers to provide the best care to their patients, they need to be informed about the types and frequency of CAM use among their patients. This can be accomplished by methodologically developing CAM instruments, rigorously implementing and assessing these instruments, and then disseminating the findings to researchers and practitioners.

Introduction

A t the end of 2006, the Centers for Disease Control and Prevention (CDC) estimated that approximately 1.1 million adults and adolescents in the United States were living with human immunodeficiency virus (HIV), the virus that causes acquired immune deficiency syndrome (AIDS), and evidence suggests that this population is still growing.¹ From 2004 to 2007, for example, the estimated number of newly diagnosed cases of HIV/AIDS in the 34 states that were reported to CDC increased 15%.² However, the number of deaths of individuals with AIDS during this same time period decreased 17%,² which is, in part, due to the success of highly active antiretroviral therapy (HAART) in extending the lifespan of those infected with HIV. In spite of pharmacological advances, however, many individuals still turn to complementary and alternative medicine (CAM) to manage HIV-related symptoms and side-effects of HAART, improve HAART efficacy and/or survival and strengthen the body.^3
–5 Indeed, one recent study reported CAM utilization rates among HIV-positive populations as high as 89%.⁶ Vitamins, herbs, nutritional supplements, and spiritual/religious healing are among the most common CAM modalities employed by HIV-positive individuals.^6

–10

Assessing CAM use among HIV-positive individuals is critical, as evidence suggests that some CAM modalities can interfere with conventional treatment regimens. For example, recent evidence suggests that St. John's Wort, garlic, and vitamin C may reduce the concentrations of HAART in the blood, thus potentially lowering its effectiveness in controlling HIV viral load.^11
–13 Other evidence suggests that HIV-positive CAM users may be less likely to adhere to their conventional treatment regimens, although this literature is conflicting.^14
–16

The risk of these adverse outcomes increases further when HIV health care providers are unaware of their patients' CAM use. In fact, many providers do not know that their patients are using CAM therapies. One recent study reported that more than half of HIV-positive individuals who reported using CAM had not told their providers.⁶ One possible reason why HIV patients may underreport CAM use to providers is that providers often fail to ask their patients about whether they have used or are currently using CAM. Wynia and colleagues (1999) report that only 7% of HIV-treatment providers reported discussing CAM therapies with every new patient, while only 5% reported discussing CAM at “most” or “every” follow-up visit.¹⁷ Other recent studies report similar findings; for example, in more than 90% of cases where the provider was unaware of CAM use, HIV-positive patients reported that the provider had not inquired about their use.⁶

Health care providers may not ask patients about their CAM use in part because they are uncomfortable with the topic due to the lack of knowledge about the efficacy and frequency of CAM use.^18
–20 For example, 61% of physicians in one study reported that they do not feel sufficiently informed about CAM safety or efficacy¹⁹; more than 70% of providers in another study claimed that they had little or no knowledge about herbal remedies.¹⁸ In spite of this lack of knowledge, most providers surveyed reported wanting to receive more education about CAM modalities¹⁹ and would recommend it to a patient if they knew it was safe and effective.²⁰ Thus, the field of CAM research has a responsibility to thoroughly and rigorously assess CAM use among HIV-positive individuals and disseminate these findings so as to further educate providers about their patients' CAM use and consequently facilitate more effective doctor–patient dialogues about CAM use.

Unfortunately, the quality of CAM assessment among HIV-positive populations is seriously lacking, thereby limiting the utility of research findings for investigators and health care providers. For example, in spite of the fact that prior research suggests that many CAM therapies are culturally-^21

–24 and stage-of-disease specific,^6,25,26 most researchers continue to implement CAM instruments that are not tailored to their study populations. Including CAM therapies on a survey that are not relevant to participants' experiences may result in poor face validity, causing lower participant motivation to perform well and provide accurate responses.^27
–29 Failure to include therapies that are relevant to participants' experiences can result in CAM use estimate inaccuracy.

Furthermore, few studies report the reliability and/or validity of CAM measures. In their review of 12 studies on CAM use among patients with breast cancer, Lengacher and colleagues reported that none of these studies cited any psychometric indices assessing the reliability or validity for CAM instruments.³⁰ In a recent systematic review of CAM use instruments implemented among HIV-positive study populations, the first author (A.O.S.) found that 20% of the studies assessed the reliability and only 3% assessed the validity of the CAM instrument employed. These measurement-related limitations consequently diminish the degree to which research findings on CAM use can be compared across studies³¹ and subsequently disseminated to HIV health care providers.

Given these limitations, the purpose of the current study was to develop and evaluate a new culturally- and stage-of-disease-appropriate measure of CAM use among a population of African-American individuals with AIDS using a mixed-method design.

Materials and Methods

Overview of the research design

This mixed-methods study used the Exploratory Design–Instrument Development model,³² a combining of qualitative and quantitative approaches for the purpose of developing and/or refining a measurement tool. This type of model in which initial qualitative data inform the development of a new, culturally appropriate quantitative measure has been successful in prior research.^33

–36

Data were collected in two phases. In phase 1, qualitative (focus group) data were used to refine an already-existing validated CAM measure for the specific study population in the present study. In phase 2, this refined quantitative instrument was then implemented in a larger sample. The resulting numeric data were analyzed to evaluate the psychometric properties of the revised CAM instrument.

Phase 1

Participants

Focus group participants were recruited from the infectious disease program (IDP) clinic of a large, public, urban hospital in the southeastern United States. Individuals were eligible to participate if they (1) were receiving their care from the IDP clinic, (2) had had an AIDS diagnosis (defined as having had a CD4+ count <200 T-lymphocytes/μL), (3) identified as African-American, (4) were 21 years of age or older, (5) spoke English, (6) were not cognitively impaired (defined as answering all questions on a brief Mini-Mental State Examination correctly³⁷), and (7) were “moderate” or “heavy” CAM users. “Moderate” or “heavy” CAM users were individuals who had used at least one type of CAM therapy in the past 12 months occasionally (some occasions per month) or on a regular basis (many occasions per week), respectively. Due to the fact that the focus groups evaluated and assisted in the refinement of an existing CAM measure, those recruited for this first phase were individuals who had experience using CAM and could therefore provide the most detailed and nuanced reflections. Such a purposive sampling technique in qualitative research is useful because it allows the researcher to access information-rich cases who can best generate the desired data.³⁸

IDP clinic patients' medical records were prescreened by the first author (A.O.S.) for eligibility. If a patient's medical record indicated that he/she had had an AIDS diagnosis, was identified as African-American, and was 21 years of age or older, a letter to his/her provider along with a recruitment flyer for the patient was inserted into his/her medical record the day before the patient was scheduled to be seen at the IDP clinic. Providers then distributed the flyers to the patient at the end of his/her clinic visit. Interested patients who called the investigator for additional information were then given the brief screening test verifying their study eligibility and, if eligible, were randomly assigned to one of the 5 focus groups.

Procedure

Following eligibility screening and enrollment and prior to participation, each participant read and signed a consent form describing the study and ensuring his/her confidentiality. All focus groups were held in a private conference room at the IDP clinic, digitally recorded, lasted approximately 60 minutes, and were conducted by the first author (A.O.S.). A research assistant (D.H.D.) was present to assist with administrative needs as well as record notes during the focus groups.

Four (4) of the five focus groups were conducted at the beginning of phase 1. Two (2) of these four groups were structured focus groups in which participants were given a paper-and-pencil version of the CAM survey used by Lengacher and colleagues,³⁰ one of the only CAM measures available that has been methodically and rigorously evaluated for reliability and validity. Originally used with patients with breast cancer, this measure was minimally edited to make it applicable for patients with AIDS. Participants were first asked to complete the survey individually. Following this individual exercise the group discussed the strengths and limitations of the original survey, whether there were any CAM therapies listed that did and/or did not seem to belong and if so, why.

The two other focus groups were unstructured in which participants were not given the edited Lengacher and colleagues³⁰ CAM survey; instead they were asked to discuss their own conceptualization of CAM without having access to an a priori framework as did the prior two focus groups. Specifically, they were asked about what comes to mind when they think about CAM, what kinds of therapies they might include in this definition, and what people use that they would consider CAM. The purpose of including two additional, unstructured focus groups was to provide a forum within which participants could define CAM on their own terms, thereby providing additional critical insight into the operationalization of CAM-related behaviors.³⁹

Both the refined CAM survey questions from the structured focus groups and domains generated from the unstructured focus groups were used to create a formative beta version of a revised CAM survey specifically tailored to the population in the present study. The fifth and final focus group completed this formative beta version using the same procedure as the prior structured focus groups except, that participants completed the revised survey on laptop computers that used an audio computer-assisted self-interview (ACASI) program (as opposed to the more traditional paper-and-pencil format). In contrast to the first two structured focus groups during which the content of the survey was refined, the purpose of the fifth and final focus group was to evaluate the readability, clarity, and face validity of the survey using a cognitive assessment technique. Therefore, the purpose of this focus group was to further evaluate the survey before it was distributed to a larger sample in phase 2. Any errors, misunderstood items/terms, or unclear elements of this version were addressed, resulting in a beta version ready to implement in phase 2. Following completion of the focus group, all individuals answered several sociodemographic questions and were compensated for their time.

Analyses

First, all digitally recorded focus group conversations and notes were transcribed. Following transcription, the resulting data were analyzed using a content analysis technique, a method “for making inferences by objectively and systematically identifying specified characteristics of messages.”⁴⁰ Specifically, this technique is most appropriate when the researcher wishes to approach narrative analysis with preidentified themes.⁴¹ All narratives were then coded for the presence of any of the following themes: (1) misunderstanding of/confusion about a survey item, (2) belief that a therapy should not be included on the survey, and (3) belief that a therapy should be included on the survey. All items that were identified as confusing or unclear were edited; all items identified as not belonging on the survey (either because they were irrelevant, upsetting, confusing, etc.) were eliminated from the survey only if identified in more than one focus group. Items identified as belonging on the survey were kept or added, even if just identified in one focus group. This approach aimed to be conservative, whereby all types of and reasons for using CAM were added more freely than they were subtracted so the instrument would be able to capture the most diverse set of experiences. Added items had the same Likert response format as those items in the original survey.

Phase 2

Participants

In contrast to the phase 1 recruitment strategy (which involved recruiting individuals who were CAM users), the phase 2 strategy focused on recruiting both users and nonusers so that the study population was more representative of the IDP clinic patient population. Eligible individuals for phase 2 (1) had not participated in phase 1, (2) were receiving their HIV/AIDS care from the IDP clinic, (3) had had an AIDS diagnosis, (4) identified as African-American, (5) were 21 years of age or older, and (6) spoke English.

Participants were recruited from the first floor of the IDP clinic, where patients are required to check-in/check-out for their appointments. The research assistant (D.H.D.) actively recruited patients by conducting preliminary screening and, if the patient were eligible to participate, referred him/her up to the third floor conference room where the survey was administered.

Procedure

Following eligibility screening and enrollment and prior to participation, each participant read and signed a consent form describing the study and ensuring his/her confidentiality. Participants then completed the CAM instrument developed during phase 1 on ACASI-programmed laptops; the instrument took approximately 30–45 minutes to complete. Following completion of the survey, all individuals answered several brief sociodemographic questions and were compensated for their time.

To conduct a test–retest reliability analysis on the stability of the CAM survey, approximately one third of the initial survey participants were randomly selected to complete the CAM survey again 2–4 weeks after completing the initial survey. This time between administrations of the survey was specifically selected because it decreased the chance that participants remembered their responses from the first administration (and answered consistently on subsequent surveys) but was not such a delay so as to risk actual changes in CAM use.⁴² Randomly selected individuals were contact via telephone, asked whether they would be willing to complete the survey again and, if they were, they returned to the same IDP clinic conference room to participate again. Test–retest participants were similarly compensated for their time.

Analyses

Following phase 2 data collection, data were entered and cleaned using SPSS version 17.0. The “frequency of CAM use” scale was then coded such that participants were assigned a “0” for each CAM therapy they reported “never” using, a “1” for each therapy they reported “seldom” using, a “2” for each therapy they reported “occasionally” using, and a “3” for each therapy they reported using on a “regular basis.” Survey data were then analyzed to assess the reliability of the “frequency of CAM use” scale using two methods. First, an estimate of internal consistency using Cronbach's α was calculated. This method provided an indication of the interrelatedness of the items. Second, Pearson correlations were calculated on the responses from those participants who completed the CAM survey on two occasions to examine the scale's test–retest reliability.

Data were also analyzed to assess the validity of the “frequency of CAM use” scale. Specifically, in order to (1) evaluate whether there are any underlying dimensions and/or subscales and (2) identify any weak survey items, an exploratory factor analysis was conducted using a maximum likelihood common factor analysis with oblique rotation. Oblique rotation is the most appropriate method when factors are expected to be correlated as in the present analysis of CAM use.⁴² To determine the number of meaningful factors to retain, Eigenvalues of 1.0 or greater were identified and scree plots were examined. Estimates of internal consistency were computed for all resulting factors. Finally, construct validity was assessed using a known-groups approach. In this approach, participants are selected based on their membership in a group that is expected to differ on the construct of interest; if the participants' scores on the instrument differ in the hypothesized way, there is evidence of construct validity.⁴³ Given that research indicates that CAM use is consistently positively associated with female gender,^6,44
–46 there would be evidence of construct validity if female participants reported more frequent CAM use than male participants. Construct validity was assessed using a one-way analysis of variance.

Results

Phase 1

Content analysis

The five focus groups conducted had approximately 6–8 participants in each (N = 35) and were similar with respect to the sociodemographic characteristics of the participants.

Participants in both structured focus groups identified many therapies listed on the original CAM survey that they did not feel were relevant to their experiences. For example, participants in both groups did not feel that HIV-positive individuals used special diets (e.g., macrobiotic), practiced cleansing regimens (e.g., fasting, using enemas), or took health food supplements (e.g., shark cartilage, barley grass). Though several participants had heard of biofeedback, electrostimulation, and light treatments, no one knew of anyone who had ever used these therapies. Most participants had never heard of ozone, metabolic or Chelation therapy. Therefore, these therapies, in addition to several others, were eliminated from the revised survey because they were identified as irrelevant in both structured focus groups and no participants in the unstructured groups offered any conflicting information supporting the inclusion of these therapies.

Participants in the structured focus groups commented that though they felt that vitamins should be included, the examples provided on the original survey (vitamin E and selenium) were not the most commonly used vitamins by HIV-positive patients. Multivitamins, vitamin C, and calcium were the suggested examples and were therefore included on the revised survey. Similarly, herbal supplements were commonly used, though several participants felt that the revised survey should specifically include yellow root and milk thistle, which were not mentioned on the original survey (which provided Ginkgo biloba and St. John's Wort as examples).

Participants in the structured focus groups consistently agreed that vitamins, herbs, chiropractic, and acupuncture were important CAM therapies to include and this sentiment was validated by participants in the unstructured groups. Only one of the structured groups felt that aromatherapy should be included (individuals in the other structured group reported not being familiar with the therapy); however, because it was identified by one group, it was retained in the revised survey. Meditation, massage, prayer/spiritual healing, and counseling/support groups were identified as being relevant therapies in both structured groups. Interestingly, after one participant in an unstructured group identified spiritual healing as an important CAM therapy, several other participants spent some time discussing whether prayer/spiritual healing should even be categorized as CAM. This debate is certainly commonplace in the field of CAM research, and some researchers have decided to exclude prayer/spiritual healing when assessing CAM use among various populations.⁴⁷ However, this therapy was retained in the revised survey because there was some consensus that prayer/spiritual healing was practiced frequently for health-related purposes.

Participants in all four groups discussed using dietary supplements such as Boost, Ensure, and other protein drinks, likely due to the fact that many health care providers at the IDP clinic occasionally recommend them to patients. Several participants in the unstructured groups spoke at great length about the various home remedies they used on occasion such as kaolin and baking soda for indigestion, epsom salts for muscle pain, and vinegar for yeast infections. Finally, t'ai chi and marijuana were suggested as additional therapies not reflected on the original survey that several participants thought might be important to include (Table 1).

Table 1.

Results from Focus Group Content Analysis

Therapy	Eliminated	Kept but edited	Kept unchanged	Added
Acupuncture			X
Aromatherapy			X
Art therapy	X
Biofeedback	X
Chelation therapy	X
Chiropractic			X
Cleansing regimens	X
Colored light treatments	X
Counseling/support Groups			X
Dietary supplements				X
Electrostimulation	X
Guided imagery	X
Health food supplements	X
Herbal supplements		X
Home remedies				X
Humor therapy	X
Hypnosis	X
Magnetic therapy	X
Marijuana				X
Massage			X
Meditation			X
Metabolic therapy	X
Music therapy	X
Naturopathy	X
Ozone therapy	X
Prayer/spiritual healing			X
Reflexology	X
Special diets	X
T'ai chi				X
Traditional Chinese Medicine	X
Vitamins		X
Yoga			X

Cognitive assessment

Participants' feedback about the readability of the revised CAM survey was generally positive and few edits were necessary. Participants felt that the survey clearly and thoroughly assessed CAM use, providing evidence of adequate face validity, and did not feel that there were any confusing or unclear questions asked. Additional readability analyses in Microsoft Word indicated that the survey was accessible for a person with a 7th-grade reading level. Several participants did struggle, however, with using the laptop keyboard mouse, so an external mouse was provided for individuals participating in phase 2.

Phase 2

Sociodemographics

One hundred and eighty-two (182) individuals participated in the survey; 59 of these individuals also participated in the test–retest process. The mean participant age was 45.4 years (standard deviation (SD) = 6.86). Participants were predominantly male (69.8%), single (80.7), not working (83.8%), had a high school education or less (60.6%), made less than $15,000 per year (82.0%), and rented their apartment or house (40.6%; Table 2).

Table 2.

Sociodemographic Characteristics of the Survey Participants (N = 182)

Characteristic	Total N (%)
Age (yrs)
≤45	90 (49.5)
>45	92 (50.5)
Sex
Male	127 (69.8)
Female	52 (28.6)
Marital status
Single	146 (80.7)
Divorced or separated	27 (14.9)
Married or partnered	8 (4.4)
Employment status
Working	21 (11.5)
Unemployed	57 (31.1)
Disabled	96 (52.7)
Other	8 (4.4)
Education
≤High school	109 (60.6)
>High school	71 (39.4)
Yearly income
<$15,000	146 (82.0)
≥$15,000	32 (18.0)
Housing status
Rent	73 (40.6)
Own	4 (2.2)
Live with others	52 (28.9)
Homeless	51 (28.3)
On HAART
Yes	56 (30.8)
No	123 (67.6)
Missing	3 (1.6)
Log10 CD4 + cell count, cells/mm³, mean (SD)^a	2.28 (0.47)
Log10 HIV RNA, copies/mL, mean (SD)^a	2.74 (1.17)

N = 92.

HAART, highly active antiretroviral therapy; HIV, human immunodeficiency virus; SD, standard deviation.

CAM use

Approximately 94% of participants (N = 171) reported currently (within the last 12 months) using any CAM. The most common types of CAM used included prayer/spiritual healing, vitamins (multivitamins and calcium), counseling/support groups, meditation, and dietary supplements (Boost/Ensure, energy drinks, protein shakes). The least common types of CAM used included acupuncture, t'ai chi, yoga, and chiropractic (Table 3).

Table 3.

Frequencies and Item Correlations of Complementary and Alternative Medicine (CAM) Therapies

	Frequency of CAM use (%)
Therapy	Never	Seldom	Occasionally	Regular basis	Item correlations with total score
Vitamins	46.3	1.4	4.8	47.6	0.425*
Herbs	91.2	0.7	4.1	4.1	0.358*
Home remedies	74.1	7.5	11.6	6.8	0.435*
Chiropractic	91.2	3.4	2.0	3.4	0.483*
Massage	88.4	2.0	7.5	2.0	0.509*
Supplements	46.9	8.2	19.7	24.5	0.475*
Aromatherapy	87.1	0.7	5.4	6.8	0.556*
Marijuana	90.5	2.7	4.8	2.0	0.236*
Counseling	35.4	3.4	25.9	35.4	0.481*
Meditation	63.3	1.4	10.2	25.2	0.556*
Yoga	93.2	1.4	2.7	2.7	0.385*
T'ai chi	95.2	1.4	2.0	1.4	0.249*
Acupuncture	98.0	0.7	0.7	0.7	0.050
Prayer	40.8	2.7	4.1	52.4	0.510*

*p < 0.01.

Reliability and validity

The means of the individual scale items ranged from 0.10 to 1.69, with SDs ranging from 0.46 to 1.45. The interitem correlations ranged from 0.15 to 0.44, suggesting that there was no redundancy among items, defined as an interitem correlation greater than 0.85.⁴² The mean interitem correlation was 0.24.

The item-to-total correlations ranged from 0.05 to 0.56, with a mean item-to-total correlation of 0.41. All items except acupuncture demonstrated high correlations with the total scale score (Table 3). Due to the fact that acupuncture was not significantly correlated with the total scale score, it was not included in subsequent analyses. The scale mean was 8.68, with a SD of 5.56. The internal consistency of the entire scale, excluding acupuncture, approached the cutoff of 0.70 for adequate reliability (α = 0.67) and the test–retest reliability was satisfactory (r = 0.79, p < 0.01).

Principal components analysis resulted in a three-factor solution with four items loading on a factor conceptualized as “home-based CAM,” four items loading on a factor conceptualized as “ingested/inhaled CAM,” and four items loading on a factor conceptualized as “body-based CAM.” Counseling loaded poorly and thus was considered an equivocal item based on the fact that none of the loadings on any of the factors was 0.30 or higher, the recommended minimum criteria for a factor loading.⁴⁸ Subsequent internal consistency analyses for each of the subscales generated by the factor analysis suggested that they have less than adequate reliability (α = 0.54 for the “home-based CAM” subscale, α = 0.56 for the “ingested/inhaled CAM” subscale, and α = 0.59 for the “body-based CAM” subscale; Table 4).

Table 4.

Factor Loadings of Complementary and Alternative Medicine (CAM) Therapies

	Factor loadings
Exploratory factors and therapies	Factor 1	Factor 2	Factor 3
Factor 1 (home-based CAM)
Home remedies	0.68	0.17	0.01
Meditation	0.66	0.38	0.14
Prayer	0.68	0.04	0.09
Aromatherapy	0.37	0.14	0.25
Factor 2 (ingested/inhaled CAM)
Vitamins	0.26	0.46	0.18
Supplements	0.20	0.72	−0.12
Herbs	0.25	0.48	−0.09
Marijuana	0.18	0.51	0.52
Factor 3 (body-based CAM)
Chiropractic	−0.14	0.08	0.72
Yoga	0.44	0.04	0.57
T'ai chi	0.29	−0.38	0.63
Massage	0.24	0.20	0.71
Equivocal
Counseling/support groups	0.26	0.28	0.27
Internal consistency^a	0.54	0.56	0.59

Estimates resulting from internal consistency analyses of each subscale consisting of the shaded items.

Results from the known-groups analyses suggest that there was marginal evidence of construct validity. Female participants did report using CAM more frequently (mean = 4.22, SD = 2.29) than male participants did (mean = 3.40, SD = 2.25), though these groups did not significantly differ [F = 2.32 (2,181), p = 0.10].

Discussion

This study presents results from the evaluation of a new instrument to assess CAM use among African-Americans with AIDS. Using a mixed-method approach, this instrument was developed to be tailored to the specific study population and evaluated for its psychometric properties, thereby addressing two measurement-related limitations found in so many prior CAM studies. The instrument had satisfactory face validity, suggesting that participants judged the survey to be an appropriate assessment of CAM use, as well as adequate test–retest reliability, suggesting that CAM use was a relatively stable phenomenon over the course of 2–4 weeks and that the instrument reliably measured this phenomenon.

Furthermore, in spite of the fact that the frequency of CAM use among this population was skewed (skewness ranged from −0.28 to 4.86), all but one survey item factored in a manner that was interpretable and generally consistent with recent theoretical conceptualizations⁴⁹ and findings from prior factor analyses.¹⁶ Home remedies, meditation, prayer, and aromatherapy all loaded together on what was characterized as a “home-based CAM” factor. These are all CAM therapies that are typically practiced in the home and cost very little. Vitamins, supplements, herbs, and marijuana all loaded together on the “ingested/inhaled CAM” factor, as these are all therapies that typically enter the body orally and are expected to produce physiological changes. Chiropractic, yoga, t'ai chi, and massage all loaded together on what was characterized as a “body-based CAM” factor, as these are therapies that involve the manipulation or movement of the body in some way and can be higher in cost than the other therapies. It is not surprising that counseling/support groups did not consistently load on one factor, as this therapy was the most negatively skewed item on the survey (skewness = −0.28) and did not logically fit any of the factors (it is seldom practiced in the home, is not ingested or inhaled, and rarely involves any manipulation or movement of the body). Perhaps if individual therapy, group therapy, and support groups had been included separately on the survey, as opposed to collapsed together as one CAM modality, these items would have formed a fourth, “mental health CAM” factor.

Unfortunately, the internal consistency reliability of each of the subscales from the factor analysis was less than adequate, which may, in part, be due to the small number of items (N = 13), the skewness of the items, or the relatively small sample size. Future research could add more therapies to each subscale or administer the instrument to a more heterogeneous sample. In the meantime, it is suggested that this instrument be used as a total scale and that the total score be used in analyses instead of scores from the individual subscales.⁴³ Additionally, there was only marginal evidence of construct validity. However, female participants did report more frequent CAM use compared to male participants; the lack of statistically significant findings is likely due to the small number of females in the sample (28.6%). Future research could assess construct validity using a similar known-groups approach with a larger sample size.

Limitations

The present study was conducted among African-Americans with AIDS who were receiving their health care from a large, public, urban hospital in the southeastern United States. Therefore, these findings concerning the prevalence and the underlying dimensions of CAM use should not be generalized to other study populations, other research settings, or other regions of the country. Due to the fact that individuals' use of various CAM modalities differ significantly according to the area of the country in which they reside, their socioeconomic status, race/ethnicity, age, and health condition, to name but a few, researchers should tailor CAM measures according to their specific study population of interest. This can be achieved using a variety of approaches, though the authors felt that using a mixed-method design in the present study was a useful and thorough way to accomplish this task. Finally, additional psychometric analyses are needed to further examine the underlying dimensions of CAM use and could include both exploratory and confirmatory factor analyses and item analysis using item response theory.⁴³

Conclusions

This research addressed several important gaps in the literature, including a lack of culturally- and stage-of-disease-specific CAM instruments and an absence of psychometric assessments of these instruments. For HIV health care providers to provide the best care possible to their patients, they need to be informed about the types and frequency of CAM use among their patient population. This can be accomplished by methodically developing CAM instruments, rigorously implementing and assessing these instruments, and then consistently disseminating the findings to both researchers and practitioners.

Footnotes

Acknowledgments

This project was supported by grant number F31AT004553 from the National Center for Complementary & Alternative Medicine.

Individuals interested in using the instrument developed in the present study should contact the first author at aowensm@emory.edu

Disclosure Statement

No competing financial interests exist.

The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Center for Complementary & Alternative Medicine or the National Institutes of Health.

References

Centers for Disease Control and Prevention. HIV prevalence estimates—United States, 2006. Morbid Mortal Weekly Rep, 2008; 57:1073–1076.

Centers for Disease Control and Prevention. HIV/AIDS Surveillance Report, 2007, 19. Atlanta: Department of Health and Human Services, 2009.

Langewitz

, Ruttimann

, Laifer

et al. The integration of alternative treatment modalities in HIV infection–the patient's perspective. J Psychosom Res, 1994; 38:687–693.

Leonard

, Huff

, Merryweather

et al. Knowledge of safety and herb–drug interations amongst HIV + individu-als: a focus group study. Can J Clin Pharmacol, 2004; 11:e227–e231.

Mills

, Wu

, Ernst

. Complementary therapies for the treatment of HIV: In search of the evidence. Int J STD AIDS, 2005; 16:395–403.

Furler

, Einarson

, Walmsley

et al. Use of complementary and alternative medicine by HIV-infected outpatients in Ontario, Canada. AIDS Patient Care STDS, 2003; 17:155–168.

Agnoletto

, Chiaffarino

, Nasta

et al. Reasons for complementary therapies and characteristics of users among HIV-infected people. Int J STD AIDS, 2003; 14:482–486.

Anderson

, O'Connor

, MacGregor

, Schwartz

. Patient use and assessment of conventional and alternative therapies for HIV infection and AIDS. AIDS, 1993; 7:561–565.

Burg

, Uphold

, Findley

, Reid

. Complementary and alternative medicine use among HIV-infected patients attending three outpatient clinics in the Southeastern United States. Int J STD AIDS, 2005; 16:112–116.

10.

Mikhail

, DiClemente

, Person

et al. Association of complementary and alternative medicines with HIV clinical disease among a cohort of women living with HIV/AIDS. J Acquir Immune Defic Syndr, 2004; 37:1415–1422.

11.

de Maat

, Hoetelmans

, Mathôt

et al. Drug interaction between St John's Wort and nevirapine. Aids, 2001; 15:420–421.

12.

Piscitelli

, Burstein

, Welden

et al. The effect of garlic supplements on the pharmacokinetics of saquinavir. Clin Infect Dis, 2002; 34:234–238.

13.

Slain

, Amsden

, Khakoo

et al. Effect of high-dose vitamin C on the steady-state pharmacokinetics of the protease inhibitor indinavir in healthy volunteers. Pharmacotherapy, 2005; 25:165–170.

14.

Jernewall

, Zea

, Reisen

, Poppen

. Complementary and alternative medicine and adherence to care among HIV-positive Latino gay and bisexual men. AIDS Care, 2005; 17:601–609.

15.

Knippels

, Weiss

. Use of alternative medicine in a sample of HIV-positive gay men: An exploratory study of prevalence and user characteristics. AIDS Care, 2000; 12:435–446.

16.

Owen-Smith

, Diclemente

, Wingood

. Complementary and alternative medicine use decreases adherence to HAART in HIV-positive women. AIDS Care, 2007; 19:589–593.

17.

Wynia

, Eisenberg

, Wilson

. Physician–patient communication about complementary and alternative medical therapies: A survey of physicians caring for patients with human immunodeficiency virus infection. J Altern Complement Med, 1999; 5:447–456.

18.

Giveon

, Liberman

, Klang

, Kahan

. A survey of primary care physicians' perceptions of their patients' use of complementary medicine. Complement Ther Med, 2003; 11:254–260.

19.

Milden

, Stokols

. Physicians' attitudes and practices regarding complementary and alternative medicine. Behav Med, 2004; 30:73–82.

20.

Winslow

, Shapiro

. Physicians want education about complementary and alternative medicine to enhance communication with their patients. Arch Intern Med, 2002; 162:1176–1181.

21.

Hsiao

, Wong

, Goldstein

et al. Variation in complementary and alternative medicine (CAM) use across racial/ethnic groups and the development of ethnic-specific measures of CAM use. J Altern Complement Med, 2006; 12:281–290.

22.

Kakai

, Maskarinec

, Shumay

et al. Ethnic differences in choices of health information by cancer patients using complementary and alternative medicine: An exploratory study with correspondence analysis. Soc Sci Med, 2003; 56:851–862.

23.

McLaughlin

, Braun

. Asian and Pacific Islander cultural values: Considerations for health care decision making. Health Soc Work, 1998; 23:116–126.

24.

Murguia

, Peterson

, Zea

. Use and implications of ethnomedical health care approaches among Central American immigrants. Health Soc Work, 2003; 28:43–51.

25.

Dwyer

, Salvato-Schille

, Coulston

et al. The use of unconventional remedies among HIV-positive men living in California. J Assoc Nurses AIDS Care, 1995; 6:17–28.

26.

Fairfield

, Eisenberg

, Davis

et al. Patterns of use, expenditures, and perceived efficacy of complementary and alternative therapies in HIV-infected patients. Arch Intern Med, 1998; 158:2257–2264.

27.

Cascio

. Applied Psychology in Personnel Management, 3rd. Englewood Cliffs, NJ: Prentice-Hall, 1987.

28.

Chan

, Schmitt

, DeShon

et al. Reactions to cognitive ability tests: The relationship between race, test performance, face validity perceptions and test-taking motivation. J Appl Psychol, 1997; 82:300–310.

29.

Robertson

, Kandola

. Work sample test: Validity, adverse impact and applicant reaction. J Occup Psychol, 1982; 55:171–183.

30.

Lengacher

, Bennett

, Kipp

et al. Design and testing of the use of a complementary and alternative therapies survey in women with breast cancer. Oncol Nurs Forum, 2003; 30:811–821.

31.

Pinkerton

, Holtgrave

, Leviton

et al. Toward a standard sexual behavior data set for HIV prevention evaluation. Am J Health Behav, 1998; 22:259–266.

32.

Creswell

, Plano Clark

, Gutmann

, Hanson

. Advanced mixed methods research designs. Tashakkori

, Teddlie

. Handbook of Mixed Methods in Social and Behavioral Research. Thousand Oaks, CA: Sage, 2003; 209–240.

33.

Kasprzyk

, Montano

, Fishbein

. Application of an integrated behavioral model to predict condom use: A prospective study among high HIV risk groups. J Appl Soc Psychol, 1998; 28:1557–1583.

34.

Montano

, Kasprzyk

. The theory of reasoned action and the theory of planned behavior. In: Glanz K, Rimer BK, Lewis FM, eds. Health Behavior and Health Education: Theory, Research and Practice, 3rd. San Francisco: John Wiley and Sons, 2002.

35.

Montano

, Kasprzyk

, von Haeften

, Fishbein

. Toward an understanding of condom use behaviors: A theoretical and methodological overview of Project SAFER. Psychol Health Med, 2001; 6:139–150.

36.

Nassar-McMillan

, Borders

. Use of focus groups in survey item development. Qual Rep, 2002; 7:1–10.

37.

Callahan

, Unverzagt

, Hui

et al. Six-item screener to identify cognitive impairment among potential subjects for clinical research. Med Care, 2002; 40:771–781.

38.

Borkan

, Reis

, Hermoni

, Biderman

. Talking about the pain: A patient-centered study of low back pain in primary care. Soc Sci Med, 1995; 40:977–988.

39.

Maxwell

. Qualitative Research Design, 41 2nd. Thousand Oaks, CA: Sage Publications, 2005.

40.

Holsti

. Content Analysis for the Social Sciences and Humanities. Reading, MA: Addison-Wesley, 1969.

41.

Bazeley

. Computerized data analysis for mixed methods research. Tashakkori

, Teddlie

. Handbook of Mixed Methods in Social and Behavioral Research. Thousand Oaks, CA: Sage Publications, 2003.

42.

Nunnally

, Bernstein

. Psychometric Theory, 3rd. New York: McGraw-Hill, 1994.

43.

DiIorio

, Dudley

, Wang

et al. Measurement of parenting self-efficacy and outcome expectancy related to discussions about sex. J Nurs Meas, 2001; 9:135–149.

44.

Bausell

, Lee

, Berman

. Demographic and health-related correlates to visits to complementary and alternative medical providers. Med Care, 2001; 39:190–196.

45.

Conboy

, Patel

, Kaptchuk

et al. Sociodemographic determinants of the utilization of specific types of complementary and alternative medicine: An analysis based on a nationally representative survey sample. J Altern Complement Med, 2005; 11:977–994.

46.

Kirksey

, Goodroad

, Kemppainen

et al. Complementary therapy use in persons with HIV/AIDS. J Holist Nurs, 2002; 20:264–278.

47.

Kristoffersen

, Fonnebo

, Norheim

. Do cancer patients with a poor prognosis use complementary and alternative medicine more often than others? J Altern Complement Med, 2009; 15:35–40.

48.

Carmines

, Zeller

. Reliability and Validity Assessment. Beverly Hills: Sage Publications, 1979.

49.

Fouladbakhsh

, Stommel

. Using the behavioral model for complementary and alternative medicine: The CAM healthcare model. J Complement Integr Med, 2007; 4:1–19.