AIDS and Behavior

, Volume 11, Supplement 1, pp 39–47

Provider-delivered, Theory-based, Individualized Prevention Interventions for HIV Positive Adults Receiving HIV Comprehensive Care

Authors

    • Department of Health Behavior, School of Public HealthUniversity of Alabama at Birmingham
    • Division of Infectious Diseases, Department of Medicine, School of MedicineUniversity of Alabama at Birmingham
  • Laura H. Bachmann
    • Division of Infectious Diseases, Department of Medicine, School of MedicineUniversity of Alabama at Birmingham
    • Department of Epidemiology, School of Public HealthUniversity of Alabama at Birmingham
    • Birmingham Veterans Administration Medical Center
  • Mollie W. Jenckes
    • Division of Infectious Diseases, Department of MedicineJohns Hopkins University, School of Medicine
  • Emily J. Erbelding
    • Division of Infectious Diseases, Department of MedicineJohns Hopkins University, School of Medicine
Original Paper

DOI: 10.1007/s10461-006-9196-1

Cite this article as:
Grimley, D.M., Bachmann, L.H., Jenckes, M.W. et al. AIDS Behav (2007) 11: 39. doi:10.1007/s10461-006-9196-1

Abstract

HIV prevention efforts are often difficult to emphasize in settings delivering comprehensive HIV care due to factors such as time constraints and differing priorities about the use of clinical time. To assist clinicians within dedicated HIV clinics to offer prevention strategies, investigators at two universities in the United States (Johns Hopkins University and the University of Alabama at Birmingham) have developed and implemented similar, audio-computerized-assisted, self-interviewing systems that have been programmed to assess individual patient risk factors and identify based on the patient’s self-assessment, the patient’s behavioral stage or, readiness for changing, each identified target behavior. Following the assessment, the systems provide printouts of key elements of this information along with individualized, theory-based intervention strategies to the medical provider. This paper will describe our efforts in developing provider-delivered, individualized, stage-based interventions intended to reduce high-risk behaviors among HIV-infected persons.

Keywords

HIV-positiveComputerized interventionsProvider-delivered interventionsHIV primary care

Introduction

In face of continuing high rates of new HIV infection in US there is a national initiative to integrate prevention into the care of known HIV positive individuals (Prevention with Positives). HIV prevention efforts are often difficult for clinicians to emphasize in settings delivering comprehensive HIV care due to time constraints and competing clinical care priorities. To assist clinicians with providing prevention efforts within dedicated HIV clinics, investigators at Johns Hopkins University (JHU) and the University of Alabama at Birmingham (UAB) have developed and implemented similar audio-computerized-assisted, self-interviewing systems that have been programmed to assess individual’s risk factors and identify a person’s behavioral stage, or readiness, for changing target behaviors described later in this paper. These systems provide brief, printed, individualized, theory-based counseling advice based on this information. The computerized risk assessments and prevention intervention messages generated (based on each individual’s responses to the behavioral assessment) are intended for use by medical providers within comprehensive HIV care settings as part of standard-of-care.

This paper describes the development of provider-delivered, individualized, stage-based interventions intended to reduce high-risk behaviors among HIV-infected persons. Because the intervention studies are ongoing, the efficacy of each intervention is not yet known. However, we believe that sharing the design of the systems with others who are serving people living with HIV/AIDS may assist with developing similar intervention systems.

We will present the rationale for our approach, describe the clinic sites and the populations being targeted, including the risk behaviors under investigation, and conclude with an explanation of the development and implementation process of the computerized intervention systems. We will also compare and contrast aspects of the two intervention models.

Rationale for Provider-delivered Interventions

JHU and the UAB investigators determined that prevention interventions were best delivered by patients’ providers for a number of reasons. JHU previously surveyed providers who indicated that they held the major responsibility for patients’ well-being and rated prevention highly. In addition, clinics at both sites are structured around a primary care model, and other professional or peer-delivered prevention messages would have required an “extra stop” after the clinical visit for the patients in an already demanding clinical environment. Moreover, the most likely approach to prevention messages becoming ‘standard-of-care’ for patients would be to incorporate theses interventions into the patient-provider contact time.

The UAB rationale was similar and was also based on empirical evidence demonstrating that healthcare providers play an important role in presenting patients with health information, helping patients adhere to treatment plans, motivating them to modify, or adopt, other health behaviors (Fink, Elliott, TSAI, & Beck, 2005; Marcus et al., 1997; Ockene, 1999).

Patients tend to view their providers as having a great deal of influence over their health-related behaviors (Goldstein et al., 1997). Moreover, the effectiveness of minimal provider-delivered interventions in primary care practice has been established for cancer and cardiovascular disease-related behaviors such as diet (Campbell et al., 1994), exercise (Marcus et al., 1997), and smoking cessation (Kottke, Battista, DeFriese, & Brekke, 1988) and holds promise for the reduction of risky sexual and substance abuse behaviors. In developing these initiatives we realized that providers operate under significant time constraints in providing HIV primary care. They also may have limited knowledge of the principles of behavior change and sometimes lack confidence when counseling patients. (McPhee, Bird, Fordham, Rodnick, & Osbornm, 1991; Park, Wolfe, Gokhale, Winickoff, & Rigotti, 2005). JHU and UAB have harnessed computer technology to address some of the barriers to prevention counseling by providers including lack of time and/or counseling skills. Both sites have developed and implemented theory-based, behavioral intervention systems for use with a wide range of patients participating in self-risk assessment on sensitive sexual and drug issues. The systems can analyze the data in seconds and generate brief counseling advice for providers to deliver during the regularly scheduled medical visit. If proven effective, these individualized, computerized intervention systems, designed to reduce risky sexual exposure among HIV-positive persons, can be replicable, sustainable and cost-effective.

Performance Sites and Patients’ Characteristics

Both sites are university associated clinics providing comprehensive HIV care. JHU AIDS Service (JHUAS) is centrally located at the Moore Clinic at Johns Hopkins Hospital, and providers service the Greenspring Station in Baltimore County and county health departments surrounding Baltimore. UAB is testing their system at the UAB 1917 Clinic an affiliate of UAB Medical Center. Both sites have clinics that are either supported by Ryan White funding for the uninsured or third party payers. Faculty at both sites conducts research and may be involved in clinical research and/or drug or vaccine studies. Each clinic is briefly described below. Characteristics of the HIV positive patients seen within each clinic are presented in Table 1.
Table 1

Characteristics of patients at the JHU and UAB clinics

Characteristic

JHU

JHU

UAB 1917 clinic

county

greenspring

(= 1,257)

(= 496)

(= 298)

 

Age

Mean age /years

41

41

40

Q1, Q3 age (years)

37, 47

37, 48

36, 47

Gender

Male

305 (61%)

237 (80%)

952 (76%)

Female

190 (38%)

61 (20%)

305 (24%)

Race/Ethnicity

White

219 (44%)

199 (67%)

660 (52%)

Black

253 (51%)

86 (29%)

597 (48%)

Asian

4 (1%)

1 (<1%)

0 (0%)

More than 1 Race

16 (3%)

10 (3%)

24 (2%)

Unknown/Unreported

4 (1%)

2 (<1%)

0 (0%)

Numbers representative of unduplicated patients seeking any clinical service within the clinics. JHU data for patients enrolled in clinics during 2003

JHU

JHU AIDS Service dates from the mid-1980s and expanded to additional counties in the 1990s to improve geographic access to quality HIV treatment for Maryland residents. Full time JHU staff travel to the counties to provide HIV/AIDS services. Medical doctors, certified nurse practitioners and physician assistants have similar responsibilities regarding HIV/AIDS care.

UAB

UAB 1917 Clinic opened in 1988 and provides comprehensive, primary care to the HIV-infected population of Alabama as well as Mississippi, Arkansas, and Tennessee.

The clinic is staffed by attending physicians (infectious diseases faculty), two nurse practitioners, a physician assistant and fellows that rotate through during their two to three year infectious diseases fellowships at the Department of Medicine.

Conceptual Framework Underlying the Intervention Systems

Client-centered counseling to assess readiness for behavior change has been accepted as a tenet of HIV-prevention counseling (US Department of Health and Human Services Public Health Service, 1994; Mertz et al., 2000), and has been shown to be effective at the individual level for preventing STDs. (Kamb et al., 1998). Interventions are more likely to be effective if they are grounded in theoretical models that organize factors influencing risk behaviors within a coherent framework (Morrison-Beedy, Carey, & Lewis., 2002). Thus, JHU and UAB have based their HIV prevention intervention for positives on the principles of behavior change from the Transtheoretical Model (TTM, also known as the “Stages of Change” Model, Prochaska & DiClemente, 1983, 1984).

TTM describes behavioral change through a continuum of stages, from precontemplation (not ready to change behavior), to contemplation, (considering making a behavior change), to preparation (ready for action), to action (adopting or quitting a behavior) and, maintenance (sustaining the behavior change over time). TTM recognizes that people in the process of behavioral change need to receive interventions tailored to their own stage in the process. In other words, once a stage has been determined, each individual needs to be provided with intervention messages targeted to her/his specific needs. Other constructs of the model (decisional balance (Prochaska et al., 1994), self-efficacy; (Bandura, 1977) and the processes of change (Prochaska, Velicer, DiClemente, & Fava, 1988) are not assessed by the automated systems but are included in the development of intervention messages. These latter constructs of the model have been described in detail elsewhere (Grimley, Prochaska, Velicer, Blais, & DiClemente, 1994; Prochaska, Diclemente, & Norcross, 1992; Velicer, DiClemente, Prochaska, & Brandenburg, 1985).

Many participating providers are familiar with the stages of change construct as applied to other health behaviors such as smoking cessation (Prochaska et al., 1988), increasing physical activity (Marcus et al., 1992), and medication adherence (Johnson, Grimley, & Prochaska, 1998). A number of our medical providers have stated that this model is appealing because it uses objective criteria and gives a (behavioral) diagnosis, modeling the approach they are familiar with.

Study Designs

JHU is conducting a controlled trial randomized at the patient level across clinics. UAB is using an interrupted time series design consisting of a series of pre- and post-tests conducted before and after the introduction of the intervention within the clinic. Both sites target English speaking, HIV+ adults (> 18 years) in continuing care. The JHU system, Computer Assisted Risk Assessment (CARA), targets all adults while the UAB system, Providers Advocating for Sexual Health INitiative (PASHIN), focuses on men who have sex with men (MSM). At both sites, patients are approached to participate when they come into the clinic for their regularly scheduled three-month follow-up appointments; however, JHU also recruits new patients. UAB is recruiting MSMs who have been sexually active in the past 6 months, so a rapid screener is conducted to determine eligibility. Other eligibility criteria for the sites include continuing care at the clinic and willingness to participate (12 months at JHU and 18 months for UAB). Both studies have been approved by the local Institution Review Board (IRB). Patients at both sites receive a $15 gift card at the completion of each survey.

Once eligibility is determined and informed consent is obtained, the computer-based surveys are completed in a private setting within each clinic. Participants are presented with brief, automated, interactive instructions on how to use the pointing device (mouse), or touch screen, and the audio option is offered with a set of headphones to protect privacy and limits literacy concerns (Bellis, Grimley, & Alexander, 2002; Turner et al., 1998). The patient can prompt the system to repeat survey items as well as response options. At completion of the assessment, pre-programmed algorithms translate the patient’s responses to the stage-related items and an individualized, theory-based, intervention advice sheet prints out. The printed material is placed in the patient’s medical folder prior to the medical encounter for the providers’ use and later returned to study staff. UAB system also prints each patient’s behavioral “prescription” which the provider hands to the patient to take with him.

The use of computerized technology for risk assessment has been shown to enhance elicitation of sensitive behavioral information (Locke et al., 1992; Turner et al., 1998; Wright, Acquilino, & Supple, 1998) without placing an additional burden on providers. However, our systems also provide automated, individualized patient feedback in a systematic and standardized manner. This allows providers to efficiently deliver individualized prevention messages within a brief time frame, leading to the further expansion of clinical care activities in the primary care setting while preserving the potential benefits of provider involvement in the intervention. The feasibility of this type of intervention (risk assessment and counseling advice) with smoking cessation has been shown in busy family practices (Demers, Neale, Adams, Trembath, & Herman, 1990), highlighting its generalizability and applicability to primary care settings.

Assessment Measures that Drive the Interventions

Both sites participate in the HRSA-funded Prevention with Positives Initiative coordinated by the University of California at San Francisco and all enrolled patients complete the comprehensive attitude and behavioral survey developed by the Enhancing Prevention with Positives Evaluation Center (EPPEC). Both sites also include stage of change assessments to generate the intervention cues and messages. Targeted risk behaviors include: less than 100% condom use for all sexual activities with main (steady) and other (casual) partners, non-disclosure of HIV status to all (main and casual) sexual partners, and substance abuse. In addition, UAB’s system focuses on the reduction of multiple sex partners and there are separate depression and substance use screens in the assessment; whereas JHU does not include multiple partners but does assess two other behaviors—needle sharing and readiness to enter a treatment program for substance abuse. Time to complete each of the stage of change assessments is based on the number of risk behaviors a patient may be engaging in and, on average, takes approximately 5 to 7 min to complete.

Intervention Development

The JHU system, CARA, was created using Microsoft Visual Basic™. The audio portion was recorded by a trained staff member. It is delivered via a standard personal computer (PC). The UAB system, PASHIN, was created within a Macromedia Authorware™ software environment (Version 7.2) using Sound Forge audio editing and is also packaged for delivery via PC. The two systems can be CD-Rom or Internet-based demonstrating the potential feasibility for wide dissemination.

Both JHU and UAB are intervening on multiple behavioral risk factors. Yet, when targeting multiple risk behaviors the question remains, “Should we intervene simultaneously or sequentially?” For example, JHU’s automated system prints out addresses the selected behaviors (disclosure, condom use, substance abuse), with disclosure and condom use addressed by partner type (main or casual). Providers have been trained to confirm the print-out information (stage of change) by re-assessing the patient. Providers may counsel on a single behavior or address two or three behaviors simultaneously per clinic visit, based on their knowledge of the patient and time available. The print-out also provides an overview of intervention approaches for each stage. A sample of the JHU’s CARA print-out is presented in Fig. 1.
https://static-content.springer.com/image/art%3A10.1007%2Fs10461-006-9196-1/MediaObjects/10461_2006_9196_Fig1_HTML.gif
Fig. 1

Sample: computer print-out from the JHU CARA program

UAB investigators, on the other hand, have developed ‘decision rules’ that determine which of the target behaviors needs to be addressed during each clinic visit. These decision rules are based on theoretical and empirical considerations. For instance, if a patient is engaging in only one target risk behavior, then the choice of which behavior to intervene upon is clear. If a participant is engaging in multiple risk behaviors, then the behavior that the person is further (or furthest) along in the stages of change is prioritized. The rationale behind this decision is that individuals who are more motivated to change a given behavior at baseline have better intervention outcomes at long-term follow-up (Prochaska et al., 1992). If a patient is engaging in two or more risk behaviors and happens to be at the same stage for each, priority decision rules are followed, based on level of risk and potential impact of each behavior for the patient as opposed to a more public health approach based on a patient’s altruism. The decision rules for same-stage behaviors have been programmed into the computer-based system and at the completion of each participant’s assessment the target behavior and stage are calculated by the computer. At each patient’s follow-up visits (every 3 months), the system “looks back” at the data from the previous visit and continues to work on that particular behavior until the patient reaches the maintenance stage. At this point, the intervention’s algorithms will determine if another risk behavior needs to be addressed, the stage that the patient is in, and a new behavioral intervention print-out is provided. In other words, the UAB’ intervention system takes a sequential approach to behavior change. The UAB system has 54 staging tables with 270 possible combinations of intervention messages based on risk behavior, partner gender, partner type, and partner’s HIV serostatus. A sample print-out from UAB’s PASHIN system is shown in Fig. 2.
https://static-content.springer.com/image/art%3A10.1007%2Fs10461-006-9196-1/MediaObjects/10461_2006_9196_Fig2_HTML.gif
Fig. 2

Sample of the UAB’s PASHIN program provider print-out for the contemplation stage at baseline: consistent condom use for insertive anal sex with casual, male partners with unknown serostatus

Briefly stated, if a patient is in the precontemplation stage for a given behavior (i.e., disclosing HIV positive serostatus, using condoms consistently, reducing the number of sexual partners, or going into substance abuse treatment) and has no intention to change, the provider interventions include messages that entail increasing the patient’s awareness of the seriousness of the problem, increasing the benefits of changing such as reducing the impact their behaviors have on others (potential of transmitting HIV to others), and thinking more about personal core values. Once a patient has moved to the contemplation stage and is thinking about change, provider messages continue to increase the patient’s knowledge about the given behavior and emphasize the positive aspects of change, counter the patient’s perceived negative outcomes associated with change and start to build the patient’s sense of self-efficacy for changing unhealthy behaviors in a variety of high-risk situations. In the preparation stage when the patient is ready for change, intervention messages stress the importance of coming up with a clear plan of action ,as well as a contingency plan for when things do not go as planned. Increasing self-efficacy and social support are hallmarks of the preparation stage of change. Once a patient has initiated change (i.e., reduces the number of sexual partners, uses condoms every time for a specific sexual activity, or discloses HIV positive serostatus to all partners) messages continue to emphasize the need for social support, strengthens self-efficacy and, in some cases, assist the patient with tactics to help avoid specific situations (people, places, and things) that may serve as “triggers” which may cause him to slip back to unhealthy behaviors. Finally, when the patient has sustained positive behavior change over time, intervention messages continue to help the patient to prevent relapse, encourage him/her to seek social support, become a role model for other HIV-positive individuals, and continue to support the patient’s positive change(s).

Provider Training

Before the introduction of our prevention studies, standard practice for JHU clinicians was to review transmission behaviors at the initial intake and then address as they saw fit at follow-up clinical visits. At UAB there were a handful of providers who were systematically addressing sex-related risks prior to the intervention implementation; however, most providers were not addressing such behaviors. Providers at both sites were recruited into the intervention studies. At UAB, all providers who agreed to participate provided written informed consent. The provider recruitment rate at both sites was 100%. The initial training of the ten clinicians recruited at JHU involved a 3-h dinner session, including a didactic presentation by an expert on stage-based interventions, followed by questions and answers and role play supervised by locally trained, resource staff. Follow-up sessions are provided for individuals or groups.

At UAB, training for 12 providers started with a 1-h didactic training including an overview of the rationale and importance of the intervention, an explanation of the theoretical framework underlying the intervention messages, and study logistics. The introduction session was followed-up by 5 h of interactive training emphasizing communication skills, with exercises on staging and role playing among the providers.

The provider advice sheets differ between the two sites. As noted above, JHU providers receive the patient’s stage of change for each risk behavior (disclosure to and condom use with ‘every partner every time’, and substance abuse) and are asked to confirm this assessment with the patient by re-assessing stage, select one or more behaviors to counsel on at the session, provide an intervention (guidance to appropriate intervention messages for each stage are provided on the sheet), and document the interaction in the standard electronic patient record (EPR).

UAB’s intervention messages consist of specific pre-written, stage-matched scenarios for a target behavior for the provider to choose from, placing less burden on the provider. These messages are stored in files within the system. Documentation of the interaction is recorded in the structure of a Provider Assessment of Intervention Session Form. This form details the time spent on the intervention, the provider’s perception of the quality of the intervention, and the patient’s response to the session. Although the computer does much of the work, providers need to be trained on how to bring up sensitive topics in a comfortable manner and to communicate effectively with their patients.

Quality Assurance

At JHU, the EPR is reviewed for each study patient to identify the risk factor assessed, the intervention provided, and the minutes spent (dose). If the information is not documented or is incomplete, the program manager contacts the provider and asks for details. Twenty percent of the patient sample at JHU complete a brief (four-item) pen-and-pencil survey on their counseling experience immediately after the session and 100% are contacted via the telephone at the completion of the program. The telephone survey assesses disclosure, sexual activity, condom use, drug use and social desirability.

At UAB, 100% of the patients are surveyed immediately after each counseling session. The UAB exit survey addresses length of intervention session, topics covered and quality of the intervention. These data are also compared to data obtained through the Provider Assessment of Intervention Session Form for cross-validation purposes. On site Quality assurance for providers is available at the clinics from the program managers.

Lessons Learned

The experience of developing and implementing provider-delivered interventions has led to several “lessons learned”. First, providers should be brought on board during the planning, development and implementation stages of the intervention. This action may give providers a sense of ownership of the intervention and promotes their thoughtful input. Second, implementing a provider-delivered intervention (even one designed to “fit” with the intention of saving clinicians’ time) results in some disruption of clinic flow and requires adjustment on the part of providers, staff and patients. For example, at UAB a participant was working on the intervention’s survey and his clinician was ready to see him. He asked the patient to stop what he was doing so the clinical care portion of the visit could be conducted. Needless to say, we lost the participant’s data for that visit. Therefore, conducting behavioral interventions within the clinic setting requires open lines of communication between involved parties to preserve harmony. Both JHU and UAB have worked diligently to build strong relationships with their providers and project staff and have made some adjustments to their protocols such as “No patients enrolled after 4 pm” to ensure that all study aspects are tightly interfaced into the normal clinic flow. The change in the clinic is generally accepted (and even embraced) over time. One example is that providers have been very clear during follow-up interviews that one of the primary functions of the intervention is to remind them to address risk behaviors with their patients. Another case is that a few UAB providers have commented that some of their patients are reporting behavioral changes; however, not all providers are willing to accept self-report data. A few ambivalent providers have stated that the real “buy in” will only come after seeing the objective data (lower STD rates). Third, training providers, even for very limited periods of time, remains challenging—both to find the time for the initial training and maintaining quality control. Therefore, interventions such as the ones described here, requiring minimal provider training and time may be more readily introduced and maintained than interventions requiring more intensive provider training.

Conclusion

The goal of both sites is to broaden the standard of HIV continuing care to include behavioral interventions proven successful with a number of other health-related behaviors such as smoking cessation, healthy eating, and increasing physical activity. The underlying theoretical model may actually be a ‘selling point’ with clinicians because it uses objective criteria and gives a (behavioral) diagnosis, modeling the approach they are familiar with. If found effective, these types of interventions may be suitable to be delivered in a cost-effective manner and expanded into a variety of HIV primary care settings both domestically and internationally.

Copyright information

© Springer Science+Business Media, LLC 2006