Predictors of medication adherence in an AIDS clinical trial: patient and clinician perceptions.
This article presents data from an AIDS clinical trial that
evaluated 238 (60 percent nonwhite) patients infected with HIV and their
clinician's perceptions of medication adherence and visit
attendance in relationship to lifestyle, psychosocial, and health belief
model (HBM) variables. Twelve sites collected data via a prospective,
multisite observational study design involving a companion study to a
larger randomized clinical trial. Baseline information was collected by
questionnaire and patient self-report on lifestyle; work and health-care
experiences; available support; and psychosocial issues, including the
HBM constructs. At follow-up visits, clinicians and patients graded
medication adherence using the same scale. Patients confidentially
reported follow-up information about lifestyle and answered HBM
questions. After 12 months, adherence with study visits was associated
with older age. Clinicians rated patients as having good adherence
significantly more often when those patients were older, were employed
at the time of enrollment, exhibited altruism as part of the reason for
enrolling in the clinical trial, and thought HIV was very serious.
Patients rated themselves as having good adherence significantly more
often if they were older, had family or friends who were infected with
HIV, and believed that being in the study was worth the trouble.
KEY WORDS: adherence; clinical trials; compliance; health belief model; HIV/AIDS
|Article Type:||Clinical report|
AIDS (Disease) (Care and treatment)
Patient compliance (Research)
|Author:||Cox, Lisa E.|
|Publication:||Name: Health and Social Work Publisher: National Association of Social Workers Audience: Academic; Professional Format: Magazine/Journal Subject: Health; Sociology and social work Copyright: COPYRIGHT 2009 National Association of Social Workers ISSN: 0360-7283|
|Issue:||Date: Nov, 2009 Source Volume: 34 Source Issue: 4|
|Topic:||Event Code: 310 Science & research|
|Geographic:||Geographic Scope: United States Geographic Code: 1USA United States|
In phase I and phase II trials with participants who have HIV,
close monitoring of medication adherence is usually part of the study
design. In phase III trials, however, medication and appointment
adherence may not be as tightly monitored or reported, and thus the
extent of nonadherence is often unaccounted for in the analyses and
interpretations of these clinical trials (Boudes, 1998; Freedman, 1990).
Such distinctions are relevant to social workers, who either help
researchers design clinical trims or help clients enroll in them. Social
workers are reservoirs of information, conduits of retention, and
problem solvers who facilitate the doctor-patient relationship when
clinical trial management is involved (Cox, 2002; Williams, 1999).
As a result of medical advances in the management of HIV, individuals who are infected are required to adhere to complicated regimens involving three or more medications, some with multiple doses per day. This contributes to the likelihood that patients will have difficulty following their regimens. Failure to take medications properly can result in the development of HIV resistance, posing a threat to the individual infected with HIV and to those to whom a resistant virus could be transmitted. Thus, it is imperative to elucidate the factors affecting medication adherence of patients who are infected with HIV; therefore, a major thrust of the present study was to reveal these medication compliance predictors.
The majority of studies evaluating medication adherence in patients with chronic illnesses such as hypertension, epilepsy, heart disease, or diabetes have demonstrated adherence rates of 50 percent or less among patients treated for these incurable but not necessarily fatal diseases (Kravitz et al., 1993; Sackett & Snow, 1979). Several reviews have provided extensive summaries of the existing general an d HIV-related adherence research literature (Besch, 1995; Ickovis & Weisler, 1997; O'Brien, Petrie, & Raeburn, 1992). Common nonadherent behaviors include discontinuing treatment prematurely, taking medications incorrectly, taking drug holidays, using treatments or substances not prescribed, and not attending appointments. Low adherence levels in a clinical trial may result in erroneous conclusions if participant adherence is not taken into account when the results are interpreted (Horowitz & Horowitz, 1993).
At the time of the design of the present adherence study--a companion to a clinical trial being conducted by a federally funded, community-based program--the only published study designed to assess the primary sociocultural determinants of HIV patient adherence within an experimental phase III research drug study had been conducted by Morse et al. (1991).The Morse et al. study used nurses' ratings of patients' medication adherence for ACTG 019, a double-blind, placebo-controlled trial of zidovudine in asymptomatic individuals with HIV. This study of mostly gay white men found that the more adherent patients lived farther from their study site; were less concerned with confidentiality; and were more likely to have emotional, economic, and social support.
The present adherence study's parent clinical trials program developed an observational adherence protocol to systematically evaluate the differential patterns of adherence with taking two different regimens of Bactrim to prevent an often-fatal HIV-related opportunistic infection. For five years, the adherence study evaluated patient and clinician perceptions of adherence with protocol-specific pill-taking regimens and study visit attendance in relationship to psychological variables and health belief model (HBM) premises.
To be eligible for the adherence study, patients had to be enrolled in the Pneumocystis carinii pneumoni (PCP) prophylaxis trial (PCP-TMS study) (El-Sadr et al., 1999) and be able to read English or Spanish. The PCP-TMS study compared daily double-strength trimethoprim-sulfamethoxazole (TMS) to thrice-weekly TMS for prevention or primary or secondary PCP in study patients with a CD4 cell count of less than 200 cell/[mm.sup.3]. At the 12 participating Community Programs for Clinical Research on AIDS (CPCRA) sites, enrollment in the adherence study was offered to patients as they were enrolled in the PCP-TMS study, but participation in the adherence study was not mandatory, producing a self-selected study population.
Study Design and Data Collection Tools
The adherence study was designed as a prospective, multisite, multimeasure observational study. Behavior was examined through semistructured patient interviews, clinical assessment, and a self-administered questionnaire. Follow-up measures were obtained every four months. At enrollment, trained research personnel administered a baseline data questionnaire. The baseline questionnaire gathered information regarding demographic variables, socioeconomic status, lifestyle habits, living arrangements, and forms of support available to patients in addition to responses to questions about clinical trials and HIV infection based on the HBM. The HBM posits that patients weigh the following factors in deciding to follow medical advice:perceived susceptibility to and severity of disease, barriers to treatment, and benefits of the recommended treatment. At scheduled follow-up visits, patient adherence with Bactrim, the study medicine for the parent PCP-TMS study, was determined by self-report and clinician ratings. Both patients and clinicians responded to the same three-point Likert-type scale, rating medication adherence behavior as good (patient took at least 80 percent of study medications), fair (patient took 50 percent to 79 percent), or poor (patient took less than 50 percent).
For the adherence study, patients were defined as having good medication adherence if they and their clinicians judged that they took at least 80 percent of the prescribed study medication (Bactrim) and as being nonadherent if they took less than 80 percent of their medication, corresponding to the fair and poor categories outlined for the parent study. Adherence with study visits was defined as patients being seen at any time within the two- to four-month study window defined by the PCP-TMS study. Patients were considered to have missed a visit if they did not attend the clinic so that a clinician could assess them. Adherence study participants who had their PCP-TMS study Bactrim discontinued were no longer assessed for adherence.
A patient self-report form--which captured information about lifestyle, patients' perceptions of their health-care providers, barriers to adhering to protocol requirements, and HBM premises--was completed at regular intervals. There were no personal identifiers on the self-report forms, and patients completed their forms in private and then mailed them directly to the CPCRA Statistical Center at the University of Minnesota. This procedure was adopted to allow the patients confidentiality in responding to questions evaluating their health-care providers. All questionnaires used were modified versions of interview schedules originally used by Morse et al. (1991).
Summary statistics for the baseline variables were calculated. Frequency distributions of the responses to each statement of the HBM were prepared. These responses were graphed into two categories: (1) neutral, agree, or strongly agree, which defines the case of the patient agreeing with the statement; and (2) strongly disagree or disagree, which defines the case of the patient disagreeing with the statement. Compliance at each follow-up visit at which study drugs were prescribed was coded as good or not good for both patient- and clinician-assessed compliance. Attendance at each follow-up visit was coded as attended or not attended. The percentages of patients classified as good compilers by patient and clinician were determined for selected variables, including possible responses to the HBM. Extensions of the logistic regression model for longitudinal data, generalized estimating equations (Diggle, Liang, & Zegers, 1994) were used to determine which variables were predictive of good compliance as assessed by the patient, good compliance as assessed by the clinician, and visit attendance. A two-step procedure was used to determine which variables to include in the final models. First, univariate models were run to select variables that were predictive at the [alpha] = .05 level. For the models predicting good compliance, variables that were predictive of either patient- or clinician-assessed compliance in at least one of the two models were included, regardless of their significance level. Their inclusion provides some data concerning the strength or weakness of the findings of those odds ratios.
From the PCP-TMS study, 378 of 2,625 total participants (14 percent) co-enrolled in the adherence study and completed the baseline data form. Baseline characteristics of the study population are shown in Table 1. The adherence study cohort differed significantly from the entire PCP-TMS cohort in two respects: The adherence cohort contained fewer Latinos--Hispanics (9 percent versus 14 percent), and more adherence study patients were taking antiretroviral therapy at baseline (79.4 percent versus 65.8 percent).
The demographic profile of the population in the PCP-TMS study shows that a majority of patients (60 percent) were people of color, predominantly African American; over 75 percent had at least a high school education; 38 percent were employed at the time of enrollment; and over 50 percent had used street drugs other than marijuana. Patient CD4 cell counts were less than 200, as required for enrollment in the PCP-TMS protocol.
Five questions based on the HBM were asked to determine patients' perceptions about participation in the PCP-TMS protocol. The results are shown in Table 2. Most (94 percent) agreed that being infected with HIV was very serious, and just over half thought AIDS was the worst disease one could get. A small percentage (less than 1 percent) agreed that being in the PCP-TMS study was more trouble than it was worth. Responses to the question addressing the personal health benefit of participating in a medication research study--"participating in a medication research study will prevent me from getting sicker"--were evenly distributed among those who agreed, were neutral, or disagreed.
Ratings of medication adherence by study patients and clinicians, along with attendance at follow-up visits, for selected baseline characteristics are shown in Table 3. In general, patients rated themselves as less adherent than did clinicians, and those who were in the youngest age group, had the least education, had used street drugs other than marijuana, and felt that their participation would not benefit the health of other people or that being in the study was more trouble than it was worth gave themselves the lowest adherence ratings. Clinicians gave the lowest adherence ratings to patients who felt participation would not help others, those who felt study participation was more trouble than it was worth, and those with the least education.
Adherence with Attendance at Study Visits
Using a very limited operational definition of visit attendance, it is seen that the mean number of missed visits for all study patients was 10 percent, ranging from 0 percent to 24.7 percent among the clinical trial units. Multivariate analysis (see Table 4) revealed that of the 41 baseline characteristics assessed, three were associated with visit adherence. Treatment assignment (daily versus thrice-weekly TMS) and older age were analyzed in 10-year increments and included those patients who chose to be in the study because of altruism. Previous incarceration was associated with missing follow-up visits.
The results of a multivariate analysis of baseline variables associated with medication adherence ratings by clinicians and patients for the PCP-TMS cohort are presented in Table 5. Treatment regimen (that is, daily versus thrice-weekly TMS assignment) was not related to adherence assessments. Clinician and patient assessments of medication adherence were compared at six and 12 months. At six months, 92 percent of patients were rated as having good adherence by clinicians, whereas 87 percent were rated as having good adherence by self-assessment ([kappa] = .251). At 12 months, the clinician and patient ratings for good medication adherence were 92 percent and 83 percent, respectively ([kappa] = .90).
Baseline characteristics significantly associated with a clinician rating of good medication adherence included older age, being employed at time of enrollment into the study, and choosing to participate in clinical trials because the information gained might help others. Baseline characteristics associated with poor adherence ratings were nonwhite ethnicity, use of illegal street drugs other than marijuana, and relating an attitude that AIDS is the worst disease one can get.
Patients' self-reported ratings of good medication adherence were statistically significant when associated with older age (as measured in five-year increments). Having family or friends infected with HIV was statistically significant when associated with good medication adherence for the PCP-TMS cohort. Being nonwhite and using street drugs other than marijuana were associated with self-reported poor adherence.
Baseline characteristics of clinical trial study patients that predicted either clinician- or patient-rated medication adherence included older age, an altruistic attitude, and social stability suggested by employment. Baseline characteristics associated with nonadherence included lack of social stability (for example, use of illegal drugs), being unemployed, and being less educated. The statistically significant baseline characteristics associated with missing study visits were the treatment assignment (daily versus thrice-weekly TMS), younger age, and altruistic motivation. These results are similar to the findings of previous adherence studies of other chronic diseases, in which adherence to medical treatment was influenced by psychopathology, social isolation, actual or perceived adverse medication effect, and treatments requiring behavioral change (Becker & Maiman, 1975; Blackwell, 1973; Griffith, 1990; Haynes, 1976).
Overall, patients in the adherence cohort of the PCP-TMS study were rated as having good medication adherence 86 percent of the time by clinicians and 70 percent of the time by their own self-assessment. The high rates of self-reported medication adherence, if accurate, may reflect the pill-taking behavior of individuals who choose to participant in research trials such as the present study. The findings of this study are strengthened by the agreement between the independent adherence assessments of patients and clinicians. However, there are several limitations and potential sources of sampling bias in the companion study. First, individuals who chose to participate in the parent PCP-TMS trial and the adherence study may differ from the at-large population of individuals who are infected with HIV. The small number of patients from the parent study who enrolled in the adherence study may not be representative of the PCP-TMS study cohort given that comparable information on social support, HBM constructs, mental illness, and incarceration was not collected for the parent study. In addition, the attention given to adherence in these studies may have produced a Hawthorne effect, inadvertently increasing medication adherence. Finally, though, agreement between clinicians and patients about predictors for good and not good adherence lends strength to the observational findings. Conclusions that can be drawn from this study are limited by the lack of an objective measure of adherence, though the adherence literature--especially in the field of HIV-related medication adherence--strongly supports the validity of self-reported measures of patient adherence.
Three studies on adherence to medicines for prophylaxis of opportunistic infections in HIV-infected individuals have been reported in the literature (Eldred, Wu, Chiasson, & Moore, 1998; Pekovic et al., 1998; Zachariah et al., 2001) and included evaluation of patients taking isoniazid (INH), cotrimoxazole, and various PCP prophylaxis regimens. Two of these studies are comparable in study design and size to the present adherence study.
Treatment of 2,960 Ugandans, 92 percent of whom were infected with HIV, at high risk for tuberculosis using three short-course (three- to six-month) prophylactic medication regimens was undertaken in a randomized, placebo-controlled but no-blind study (Pekovic et al., 1998). Criteria for enrollment included having a positive or anergic tuberculin skin test response, living within 20 kilometers of the project site, and willingness to be followed. Adherence was measured by testing urine for presence of INH metabolites, attendance at monthly follow-up appointments, and self-report. A compliance questionnaire given to patients at the one- and three-month visits gathered information on knowledge of regimen and amount of medication taken. During the study, home visits were made by staff trained in motivational techniques to patients who missed appointments or had negative results on the INH urine test. Analyses were done on the individual measures and on combinations of the measures after development of composite indices using all three assessment methods, resulting in a score for each patient. Eighty-six patients attended at least two-thirds of scheduled study visits. Overall, though knowledge of the regimen and dosing was high (greater than 90 percent), 30 percent of all patients reported forgetting to take some of their medication. Of the total number of urine samples tested for INH, 77 percent were positive. Analysis of individual measures of adherence showed that correlations among them were weak, that each measure provided information on a different aspect of adherence, and that they were not redundant.
Consistently across all analyses, knowledge of the regimen and features of the study arm were independently associated with adherence. The characteristics that were associated with better adherence included being older, having a better knowledge of the medication regimen, not having a military occupation, being on the simplest regimen (fewest number of pills), being skin-test positive, and having a reason for enrollment other than just to learn one's tuberculosis or HIV status.
Some of the results of the present adherence study echoed these findings in that the more adherent patients were older, and their motivation for entering into the study appeared related to later adherence. Patients who entered the Ugandan INH prophylaxis trial to resolve their personal concerns about their tuberculosis or HIV status had significantly lower scores on all composite indices than did patients who stated other reasons for participation.
A large study of adherence to PCP prophylaxis was performed at the Johns Hopkins Hospital HIV Clinic before the advent of effective combination therapy (Eldred et al., 1998). Potential study participants were recruited from a group of patients considered to be engaged in care (at least one clinic visit in the previous six months) and who had already been on antiretroviral medicines for at least six months. Eighty-five percent of the study cohort were African American, 63 percent were male, 54 percent had injecting drug use as an HIV risk factor, and 69 percent had CD4 cell counts under 200/[mm.sup.3]. Medication adherence was measured by self-report of total dose of medication taken over the previous seven days and total number of days in the previous two weeks on which medication was taken. Medical records were reviewed to derive the proportion of pills taken (number taken divided by number prescribed). To verify self-report, a urinary test for unconjugated sulfamethoxazole was performed using mass spectophotometry in a subset of those prescribed TMS who reported taking TMS in the previous three days. Adherence to pentamidine was assessed by chart review. Adherence of more than 80 percent was considered acceptable, based on previous data from studies of the treatment of latent tuberculosis.
Sixty percent of study patients took at least 80 percent of their antiretroviral medicines, and 49 percent took at least 80 percent of their prescribed PCP prophylaxis medicines. Of those reporting adherence to TMS, 70 percent had detectable urinary sulfamethoxazole, meaning they had taken this medicine within the previous 48 hours. Reasons for not taking medicines included forgetfulness (30 percent), fear of side effects (26 percent), and drug or alcohol use (16 percent).
Sociodemographic characteristics and number of prescribed medicines were not associated with either PCP or antiretroviral adherence. Associated with adherence to PCP prophylaxis by self-report and review of medical records were likelihood of taking medicines away from home, positive self-efficacy (belief in ability to adhere to therapy), presence of family, ability to afford medication copayments, and higher scores on mental health testing. Injecting drug use was associated with poor adherence to taking PCP prophylaxis therapy in multivariate analysis.
The present adherence study differed in several respects from the Johns Hopkins study. Demographically, there were fewer women, injecting drug users, and minorities in the non-Hopkins study cohort. Patients who were recruited in this cohort also included those new to HIV care and established patients. In the adherence study, information was collected on education as a surrogate for socioeconomic status, and as 20 percent were college graduates, study patients were from potentially more diverse socioeconomic backgrounds. The sample size for study patients taking TMS for PCP prophylaxis was smaller (N = 135) in the Hopkins cohort than in federally funded cohort (N = 378).
Despite these differences, the types of characteristics predictive of adherence during follow-up are confirmatory of observed data trends. These include enhancement of the social stability of patients infected with HIV by involving family or friends; provision of alcohol and substance abuse treatment programs in primary care settings; maximum use of available support from social service agencies; and, for the young, continued focused educational efforts on the course of HIV infection and its treatment. More marginal patients tend to be less adherent to medications and visits and, thus, may require research sites to invest more resources in them to maintain adherence. One final noteworthy finding is that doctors and patients were in relatively high agreement with regard to accessing perceived patterns of medication adherence.
Original manuscript received January 3, 2008
Final revision received August 7, 2008
Accepted December 3, 2008
Becker, M. H., & Maiman, L.A. (1975). Sociobehavioral determinants of compliance with health and medical care recommendations. Medical Care, 13, 10-24.
Besch, C. L. (1995). Compliance in clinical trials. Journal of Acquired Immune Deficiency Syndrome, 9, 1-10.
Blackwell, B. (1973). Drug therapy: Patient compliance. New England Journal of Medicine, 289, 249-252.
Boudes, P. (1998). Drug compliance in therapeutic trials: A review. Controlled Clinical Trials, 19, 257-268.
Cox, L. (2002). Social support, medication compliance and HIV/AIDS. Social Work in Health Care, 35, 425-460.
Diggle, R, Liang, K., & Zegers, P. (1994). Analysis of longitudinal data. Oxford, England: Clarendon Press.
Eldred, L., Wu, A.W., Chiasson, R.E., & Moore, D. (1998). Adherence to antiretroviral and pneumocystis prophylaxis in HIV disease. Journal of Acquired Immune Deficiency Syndrome and Human Retrovirology, 18, 117-125.
El-Sadr, W. M., Luskin-Hawk, R., Yurik, T. M., Walker, J., Abrams, D., John, S. L., et al. (1999). A randomized trial of daily and thrice-weekly trimethoprim-sulfamethoxazole for the prevention of Pneumocystis carinii pneuomonia in human immunodeficiency virus-infected persons. Clinical Infectious Diseases, 29, 775-783.
Freedman, L. S. (1990). The effect of partial noncompliance on the power of a clinical trial. Controlled Clinical Trials, 11, 157-168.
Griffith, S. (1990). A review of the factors associated with patient compliance and the taking of prescribed medications. British Journal of General Practice, 40, 114-116.
Haynes, P,. B. (1976). A critical review of the "determinants" of patient compliance with therapeutic regimens. In D. L. Sackett & R. B. Haynes (Eds.), Compliance with therapeutic regimens (pp. 26-39). Baltimore: Johns Hopkins University Press.
Horowitz, R. I., & Horowitz, S. M. (1993). Adherence to treatment and health outcomes. Archives of Internal Medicine, 153, 1863-1868.
Ickovis, J. R., & Weisler, A.W. (1997). Adherence in AIDS clinical trials: A framework for clinical research and clinical care. Journal of Clinical Epidemiology, 50, 385-391.
Kravitz, R. L., Hays, R. D., Sherbourne, C. D., DiMatteo, M. R., Rogers, W. H., Ordway, L., & Greenfield, S. (1993). Recall of recommendations and adherence to advice among patients with chronic medical conditions. Archives of Internal Medicine, 153, 1869-1878.
Morse, E.V., Simon, P. M., Coburn, M., Hyslop, N., Greenspan, D., & Balson, P. M. (1991). Determinants of subject compliance within an experimental anti-HIV drug protocol. Social Science and Medicine, 32, 1161-1167.
O'Brien, M. K., Petrie, R., & Raeburn, J. (1992). Adherence to medication regimens: Updating a complex medical issue. Medical Care Review, 49, 435-454.
Pekovic, V., Mayanja, H., Vjecha, M., Johnson, J., Okwera, A., Nsubuga, P., et al. (1998). Comparison of three composite compliance indices in a trial of self-administered preventive therapy for tuberculosis in HIV-infected Ugandan adults. Journal of Clinical Epidemiology, 51, 597-607.
Sackett, D. L., & Snow, J. C. (1979). The magnitude of compliance and non-compliance. In R. B. Haynes, D. Taylor, & D. L. Sackett (Eds.), Compliance in health care (pp. 11-22). Baltimore: Johns Hopkins University Press.
Williams, J. (1999).The role of social work in HIV/AIDS clinical trials. Social Work in Health Care, 29, 35-56.
Zachariah, R., Harries, A.D., Arendt,V., Wenning, R, Schneider, S., Spielmann, M., et al. (2001). Compliance with cotrimoxazole prophylaxis for the prevention of opportunistic infections in HIV-positive tuberculosis patients in Thyolo district, Malawi. International Journal of Tuberculosis Lung Disorders, 5, 843-846.
Lisa E. Cox, PhD, LCSW, MSW, is associate professor, School of Social and Behavioral Sciences, Richard Stockton College of New Jersey, P.O. Box 195, Jimmie Leeds Road, Pomona, NJ 08240; e-mail: firstname.lastname@example.org. This article was assembled with the use of data provided by the Terry Beirn Community Programs for Clinical Research on AIDS (CPCRA) (National Institutes of Health, National Institute of Allergies and Infectious Diseases, Contract Numbers Y01-AI-0002, Y01-A1-4040, and Y1-AI-0431), and the author thanks the CPCRA for its early support and willingness to share data from the CPCRA 012 and CPCRA 006 studies. In addition, gratitude and acknowledgement goes to the following original CPCRA units and members for their intellectual, statistical, and writing and editing contributions: C. Lynn Besch, MD, Glenn Bartsch, PhD, Carroll Child, PhD, Donald L Abrams, MD, the CPCRA Richmond AIDS Consortium clinical trials unit, and Betsy Finley, RN
Table 1: Summary Statistics of Selected Patient (N = 378) Characteristics at Baseline Variable M(SD) % Age (years) 39.0 (8.8) Female 13.0 Nonwhite 60.3 History of prior injecting drug use 29.6 CD4 cell count (cells/mm3) 69.0 Prior AIDS diagnosis 36.2 Education Less than high school 21.8 High school graduate 58.1 College graduate 20.2 Currently employed 37.7 Chose to take part in the study to help other people 62.3 Prior use of street drugs other than marijuana 52.2 Note: One patient had missing data for the variables indicated Table 2: Frequency Distribution of Responses to Health Belief Model Statements at Baseline Strongly Disagree Disagree Health Belief Number Model Statement Responding n % n % Being infected with HIV is very serious. 377 6 1.6 4 1.1 AIDS is the worst disease you can get. 377 17 4.5 46 12.2 Participating in a medication research study will prevent me from getting sicker 376 42 11.2 89 23.7 Participating in a medication research study will benefit the health of other people. 377 2 0.5 2 0.5 Participating in a medication research study is more trouble than it's worth. 376 151 40.2 168 44.7 Health Belief Model Statement Neutral Agree n % n % Being infected with HIV is very serious. 13 3.4 65 17.0 AIDS is the worst disease you can get. 72 19.1 84 22.0 Participating in a medication research study will prevent me from getting sicker 115 30.6 79 21.0 Participating in a medication research study will benefit the health of other people. 25 6.6 17 47.0 Participating in a medication research study is more trouble than it's worth. 47 12.5 6 1.0 Table 3: Frequency of Reporting Good Medication Adherence or Attendance at Follow-up Visits by Level of Selected Baseline Characteristics Visits at Which Bactrim Adherence Was Rated Good (%) By By Characteristic n Patient Clinician Randomized treatment assignment (dosing schedule) Daily 189 78.7 85 3x/week 189 79.1 86 Total 378 Age (years) <35 115 71.5 80 35-45 185 80.0 86 >45 78 85.1 89 Total 378 Nonwhite Yes 228 71.5 82 No 150 88.5 89 Total 377 Education Less than high school 82 64.1 77 High school 219 80.7 86 College graduate 76 87.7 92 Total 377 Currently employed Yes 142 82.6 90 No 235 76.1 82 Total 377 Chose to participate in the study to help other people Yes 256 80.1 89 No 11 77.3 78 Total 377 Ever used street drugs (other than marijuana) Yes 197 72.9 81 Total 377 Note: Yes = response of neutral, agree, or strongly agree; No = response of strongly disagree or disagree. Table 4: Baseline Characteristics Associated with Visit Adherence from the Generalized Logistic Regression Revealing the Odds Ratios of Good versus Not Good Parent Study Cohort Adherence at Follow-up Visits Assessed by Patient Characteristic OR (95% CI) p Age (10-year increment) 2.1 (1.5. 3.1) .0001 Nonwhite 0.3 (0.2, 0.7) .002 Currently employed 0.9 (0.5, 1.8) .81 Education (4-year increment) 2.6 (1.4, 4.9) .002 Chose to be in the study to help others 1.1 (0.6, 2.0) .80 Prior use of street drugs (other than marijuana) 0.5 (0.3, 0.9) .014 Agrees that being infected with HIV is very serious 1.4 (0.4, 5.4) .61 Agrees that AIDS is the worst disease you can get 1.2 (0.6, 2.4) .58 Assessed by Clinician Characteristic OR (95% CI) p Age (10-year increment) 1.5 (1.1, 2.0) .005 Nonwhite 0.7 (0.4, 1.2) .21 Currently employed 2.0 (1.1, 3.5) .013 Education (4-year increment) 1.4 (0.9, 2.3) .17 Chose to be in the study to help others 2.5 (1.5, 4.1) .0006 Prior use of street drugs (other than marijuana) 0.5 (0.3, 0.8) .005 Agrees that being infected with HIV is very serious 2.8 (1.0, 7.6) .049 Agrees that AIDS is the worst disease you can get 0.6 (0.3, 1.0) .066 Note: Good = 80 percent or greater; Not Good = less than 80 percent; OR = odds ratio; CI = confidence interval. Table 5: Baseline Characteristics Associated with Medication Adherence from the Generalized Logistic Regression Revealing the Odds Ratios of Good versus Not Good Patient Adherence at Follow-up Visits Characteristic OR (95% CI) p Treatment assignment (daily versus thrice-weekly) 0.5 (0.3, 0.9) .023 Age (l0-year increment) 1.7 (1.2, 2.5) .003 Agrees that participating in a medication research study will benefit the health of other people 3.2 (1.2, 8.7) .024 Note: Good = greater than 80 percent; Not Good =less than 80 percent; OR = odds ratio; CI = confidence interval.
|Gale Copyright:||Copyright 2009 Gale, Cengage Learning. All rights reserved.|