People living with HIV/AIDS (PLWHA) (methods)

toolkit material

Monitoring behaviour among PLWHA is a sensitive issue because of the risks of stigmatisation, criminalisation or existing legal conditions regarding transmission of HIV or other STI. Any behavioural surveillance should be discussed with PLWHA organisations. The populations to be considered are people diagnosed with HIV and treated in specialised settings for their HIV infection as well as subpopulations: women, men self identified as heterosexuals or refusing self identification, men self identified as homosexuals or bisexuals. There is no referenced sampling method to monitor trends in the population of people living with HIV. In most countries, HIV declaration is anonymous, and hence there is no registry to serve as a reference population for sampling in medical settings.

In Europe, behavioural surveillance has been conducted mainly using service-based sampling.

To date, data have been collected in one-day studies, in cohorts and in cross-sectional studies. These methods are detailed below:

  1. One-day prevalence survey 
  2. Cohort studies
  3. Cross-sectional repeated surveys
  4. Advantages and disadvantages of these methods are presented in a table

1. One-day prevalence survey

This type of survey aims at recruiting an exhaustive sample of subjects attending outpatient clinics or hospitalised on one fixed day, using a short questionnaire administered by the physician or the nurse in charge of treatment and medical care. Behavioural information is part of a questionnaire also covering medical information and the social situation. The questionnaire is completed during the visit. Data collection is initiated and organised by health state agencies. Such a survey can be performed annually. Feasibility depends on the organisation of HIV care in the country, the organisation of health administration to implement the study in all, or at least in large samples of, specialised settings. As data are collected by nurses and physicians, little infrastructure is needed. Participation of hospitals may vary across regions and years, however patient response rates are high. One-day studies will underrepresent both those patients who attend medical settings infrequently because they are in good health and have no need for treatment and those who have interrupted their treatment or do not comply to regular visits. Desirability bias in answering questions about behaviour, especially if questionnaires are more detailed than usual, might occur. This type of data collection allows only for short questionnaires, but the method is feasible and sustainable due to the limited resources needed.

2. Cohort studies

Cohort studies are mainly research tools focusing on clinical, physiopathological and therapeutic issues. To date, little interest has been given to behavioural issues. Only few cohorts collect behavioural data (e.g. the Swiss HIV Cohort Study, the French PRIMO cohort) or provide analysis of such data. This situation may change, as there is increasing interest in the link between treatment and prevention. Data collection within cohorts is facilitated by the availability of an existing research infrastructure, funding and technical assistance. However, existing cohorts are heterogeneous: some national cohorts are essentially longitudinal medical and biological databases (Swiss Cohort), others recruit a specific segment of subjects, such as individuals recently diagnosed (COPANA, France), seroconverters (CASCADE, France), patients co-infected (HEPAVIH, France); others have a local or regional basis. Cohort studies are not representative of the national population living with HIV, although some cover a large proportion of the target population or aim at exhaustivity. However, the key issue regarding use of the cohort as a surveillance tool is whether the cohort has ceased to include new cases or is continuing recruitement. Populations followed in the cohorts are selected both at inclusion and by attrition. Methods in use are either self-administered questionnaires or questions asked by the physician or nurse. The interest in second generation surveillance is the emerging with detailed medical and biological markers. Countries setting up new cohorts should be invited to include behavioural data. Validity is determined by the scope of the cohort, and comparability between countries is limited by the characteristics of patients recruited in the cohorts. Feasibility as a behavioural surveillance tool is facilitated by the fact that use is made of existing infrastructure at limited additional costs. Cohorts are long-term studies, and this also ensures sustainability.

3. Repeated cross-sectional surveys

Repeated cross-sectional surveys using either self administered questionnaires, face to face or computer-assisted confidential questionnaires can recruit large samples and collect detailed data to document indicators. Information from HIV notifications and medical setting case-loads can provide information allowing for the definition of random samples within services, allowing for a random selection of patients within the clinics. Such studies require considerable specific resources and are less likely to be sustainable. In the current situation, a brief questionnaire allowing for the collection of information necessary to document indicators could be proposed for inclusion in each type of design. Information regarding the context of HIV care and statistical data on the characteristics of case-loads across type of settings and regions are necessary to complement data collected at diagnosis on transmission categories, demographic characteristics, and progression of HIV.

Table 1 presents selected features of available methods for collecting behavioural surveillance data among PLWHA.

Table 2  presents the best two methods proposed by the experts for the situation in which the population is reachable in known settings.

Table 1. Selected features of available methods for collecting behavioural data among PLWHA

 

Design Advantages Disadvantages Comments
Population representative designs    

In countries with concentrated or low epidemics, population surveys are not adapted to surveillance among PLWHA: no nominative registers, very small numbers in the general population

 «Convenience», non-representative designs

Venue-based surveys (a) Practical, can be linked with an intervention  

Venues attended by PLWHA are medical settings or places run by NGOs which select segments of the PLWHA population 

  • Indication: qualitative surveys targeting special sub-groups, such as African migrant women etc.
  • PLWHA organisations provide services to segments of the population living with HIV and might allow for exploration of the experience of living with HIV, these issues being infrequent or difficult to report in studies targeting the whole HIV+ population
Internet surveys (b)     Practical, possibility to reach many individuals Internet allows HIV-infected MSM to be reached but will under- represent other populations, such as migrants and older patients
  • Internet has been little used 
  • Despite limitation and bias, Internet might allow for the exploration of small subgroups and of hidden behaviour, difficult to report (such as not telling status to stable sexual partners, deliberately risky or violent behaviour)

Service-based surveys, two main types (c)

In most EU countries, people with known HIV infection receive care and most receive ARV treatment in special services in either hospitals or specialised medical practices

 

Underrepresent people who interrupt treatment and care; Interviewing on beha-viour might be biased in the medical setting

Is the best place to obtain diversi¬fied samples, since most PLWHA attend medical visits on a very regular basis

Cross-sectional surveys among HIV+ individuals attending hospital settings (d)
  • Possibility of random sampling
  • A large proportion of persons with HIV consult several times a year, whatever their health, social status and transmission group 
  • Sustainable
  • Such surveys are costly if questionnaires are admini-stered by interviewers 
  • Self-administered questionnaires may limit participation of migrants 
  • Questionnaire administered by doctors/nurses might be subject to desirability bias 
  • In countries where a significant proportion of patients are cared for in private practice, such a design will not account for this segment of HIV- infected population 
  • Different designs are possible:
  • One day annual prevalence survey on an exhaustive sample of hospitals (a)Limited medical and behavioural information is collected during the visit among attending patients. Such a study is sustainable with limited costs but collects only little information
  • Repeated cross-sectional surveys using either self-administered, face to face or computer-assisted confidential questionnaires might recruit large samples and collect detailed data on indicators. Such studies are costly,since they require resources to recruit the patients on a random basis. Information is collected either by self-administered questionnaire (b) or by a trained interviewer (c)
Cohort studies (e)
  • Cohorts exist in a significant number of EU countries and document a range of clinical and medical information
  • Data collection within cohorts is facilitated by existing research infra-structure, funding and tech-nical assistance
  • Additional costs would be limited
  • HIV Cohorts are long- term studies ensuring sustainability 
  • Cohorts are heterogeneous across countries and do not account for the whole population living with HIV 
  • Cohorts are selective both at recruitment and by attrition
  • Comparability across countries might be limited 
  • Cohort studies are mainly research tools focusing on clinical, physiopathological and therapeutic issues with little interest up until now in behavioural aspects 
  • Some cohorts collect behavioural data: the Swiss Cohort), the French PRIMO cohorte)

 

Examples (a) Venue-based surveysSpain: http://www.isciii.es/htdocs/centros/epidemiologia/epi_sida.jsp

(b) Internet surveysElford J, Bolding G, Davis M, Sherr L, Hart G.  The Internet and HIV study: design and methods.  BMC Public Health 2004;4:39.

(c) Service-based surveysPeretti-Watel P, Riandey B, Dray-Spira R, et al. (2005). Comment enquêter la population séropositive en France: l'enquête ANRS-EN12-VESPA 2003 [Surveying the HIV-Positive Population in France. The ANRS-EN12-VESPA 2003 Survey]. French. Population. 60, 4, 525-550.

(d) Cross-sectional surveys among HIV+ individuals attending hospital settingsGlass TR, Young J, Vernazza PL, Rickenbach M, Weber R, Cavassini M, et al. Is unsafe sexual behaviour increasing among HIV-infected individuals AIDS 2004;18(12):1707-14. Available at: URL: PM:15280782.

Elford J, Anderson J, Bukutu C, Ibrahim F. HIV in East London: ethnicity, gender and risk (2006). Design and methods. BMC Public Health, 9, 6, 150.

Hasse B, Ledergerber B, Hirschel B, Vernazza P, Glass TR et al.. Frequency and determinants of unprotected sex among HIV-infected persons: the Swiss HIV Cohort Study. Clinical Infectious Diseases 2010,51(11):1314-22

Thurnheer MC, Weber R, Toutous-Trellu L, Cavassubu M, Elzi L. Occurrence, risk factors, diagnosis and treatment of syphilis in the prospective Swiss HIV Cohort Study. AIDS, 2010, 24(12):1907-16.

(e) Cohort studiesDesquilbet L, Deveau C, Goujard C, Hubert JB, Derouineau J, Meyer L. Increase in at-risk sexual behaviour among HIV-1-infected patients followed in the French PRIMO cohort. AIDS 2002;16(17):2329-33. Available at: URL: PM:12441806.

The table below presents the best two methods proposed by the experts for the situation in which the population is reachable in known settings.

Table 2. Best methods to access PLWHA 

 

Method Main indication for preference
Service-based cohort studies 

Most persons living with known HIV consult several times a year, whatever their social status or transmission group

Service-based cross-sectional studies 

In countries where there are no cohorts or where cohorts do not include any behavioural information