An evaluation of partner notification for HIV infection in genitourinary medicine clinics in England

Abstract

Objective: 

To evaluate the feasibility and effectiveness of a standardized HIV partner notification programme within genitourinary medicine clinics in England.

Design: 

A prospective survey of HIV partner notification activity over a 12-month period.

Setting: 

Nineteen genitourinary medicine clinics in England.

Patients and participants: 

A total of 501 eligible HIV-positive patients (either newly diagnosed or with whom partner notification had not been undertaken previously) seen during the study period.

Main outcome measures: 

The numbers of partners named by patients, and the number of contacts notified, counselled and HIV-tested.

Results: 

Information on overall partner notification activity was obtained by reviewing available medical records of 471 patients; 353 (75%) had discussed partner notification with a health-care worker during the study period and 197 (42%) had undertaken partner notification. Detailed information on outcomes was obtained for only 70 patients who named 158 contacts as being at risk of acquiring HIV. Although 71 (45%) contacts were eventually notified, only 28 were subsequently seen in participating clinics. Almost all contacts (n = 27) requested HIV counselling and testing, and five were diagnosed HIV-positive. Patient referral was the most popular notification method chosen.

Conclusions: 

This study illustrates some of the practical difficulties that limit HIV partner notification within genitourinary medicine clinics. These include health-care workers' misgivings about undertaking partner notification, insufficient locating information to identify contacts, and migration of newly diagnosed patients, which prevents continuity and completion of notification. Nevertheless, HIV partner notification uncovered previously undiagnosed HIV infections. Further work needs to be undertaken in staff training and policy implementation if higher rates of partner notification and outcome measurements are to be achieved.

Introduction

Partner notification (or contact tracing) is a strategy used in the management of patients with sexually transmitted diseases (STD) whereby sexual partners (contacts) are identified and informed of their potential risk of infection [¹]. It aims to facilitate the early diagnosis and treatment of STD, thereby curtailing its spread and protecting the health of individuals, their contacts and the wider community. Partner notification has been advocated to prevent HIV transmission alongside HIV counselling and testing, promotion of safer sex and condom use. Its use has been controversial [^2–5] and there are wide variations in HIV partner notification practice internationally [^6–9]. Ethical concerns and uncertainty regarding the feasibility and social repercussions of this strategy remain unresolved [^10,11].

The UK Department of Health issued guidance on HIV partner notification in December 1992 [¹²]. This provided detailed procedures for undertaking HIV partner notification and encouraged health authorities to develop local policies in a variety of settings including genitourinary medicine (GUM) and antenatal clinics. The guidance stressed the importance of encouraging HIV-infected patients to inform their partners about their exposure (patient referral), and that the option of provider referral (notification of partners by a health-care worker) should be offered to assist this process.

This study aimed to evaluate the feasibility and effectiveness of a standardized HIV partner notification programme when undertaken within GUM clinics. The main objectives were to determine the extent to which such a programme could succeed in identifying and locating people unaware of their increased risk of HIV infection, and the resources (health adviser time and costs) required to do so. In this article, we describe the design and implementation of the study, and present the outcomes of partner notification among newly diagnosed HIV-positive patients.

Methods

The diagnosis and management of STD and HIV is carried out in a network of over 200 GUM clinics in England. In this setting, health advisers are primarily responsible for counselling, partner notification and health education. Twenty-one GUM clinics were purposively selected to participate in this study on the basis of their HIV-positive patient caseloads (high or medium throughput), geographic locations (metropolitan, urban, sub-urban) and existing HIV partner notification practices. One clinic had already evaluated its partner notification programme and another declined to participate. The demographic characteristics of the 19 participating clinics have been published previously [¹³].

Subjects

Recruitment to the study took place over a 12-month period commencing 1 April 1994. All HIV-positive patients who were diagnosed or managed for the first time in the 19 participating clinics during the study period, and had not previously undertaken partner notification, were eligible for entry. In keeping with current practice, HIV-positive patients were to be referred to a health adviser by their attending physician for counselling. At an appropriate time in the consultation, the issues of notification, counselling and HIV testing of sexual or drug injecting contacts were brought up by the health adviser. Health advisers were also responsible for introducing the study, recruiting and interviewing eligible patients. All recruited patients (index patients) were required to give written consent to participate in the study.

Training

Health advisers in GUM clinics have varied backgrounds, including nursing, health promotion and counselling, and many clinics do not provide training on HIV partner notification methodology [¹³]. Therefore, all health advisers participating in the study were required to attend a 3-day residential training course, which aimed to establish standard policies and codes of practice for HIV partner notification. A project manual (including the study protocols) was given to health advisers in each participating clinic as a resource on HIV partner notification methodology. Participating clinics obtained ethical approval from their local committees.

Data collection

Once recruited, index patients were asked to identify their sexual or drug injecting contacts. ‘Contacts’ were defined as partners who may have been at high risk of acquiring HIV either through unprotected penetrative sexual intercourse or by sharing contaminated injecting equipment since the index patients' possible time of seroconversion (self-assessed). Notification of contacts was said to be ‘verified’ if it could be confirmed by a health adviser who was participating in the study.

Health advisers used three data collection forms throughout study. The first collected information from the index patient about themselves. The second was used to obtain information from the index patient about each of their sexual or drug injecting contacts. The third was used to collect information from notified contacts about themselves. Demographic and behavioural risk data as well as information on the time spent by health advisers discussing partner notification, contacting partners or supporting contacts were recorded on all three forms.

At the end of the study, the medical records of all eligible patients seen during the study period were reviewed to obtain data on the demographic characteristics (gender, location of diagnosis, risk category) of non-recruited patients. If documented, the outcomes of partner notification activity, as well as the reasons for non-recruitment were also collected. HIV partner notification was said to be ‘discussed’ if patients' sexual or drug injecting partners were documented in their medical records in relation to notification, and ‘undertaken’ if there was documentation of the outcomes of partner notification (e.g., partner notified, cancelled or tested).

Statistical analysis

All analyses were conducted using SPSS version 6.0 for Windows (SPSS Inc., Chicago, Illinois, USA). Proportions were compared using the χ² test.

Results

Overall partner notification activity

A total of 501 eligible HIV-positive patients were seen in the 19 participating clinics during the 1-year study period, and 471 medical records were reviewed 2 months after recruitment ended. Thirty notes were unavailable and presumed missing. Documentation of partners or partner notification was found in the records of 353 (75%) patients (Fig. 1).

Fig. 1: . Outline of HIV partner notification activity and outcomes. ^*Partners believed to be at risk of acquiring HIV since possible seroconversion through unprotected sexual intercourse or sharing contaminated injecting equipment. ^†Contacts verified if notification was confirmed by a health adviser from the project participating clinic. GUM, Genitourinary medicine.

Documentation of the outcomes of partner notification was found with 197 (42%) patients. (Fig. 1). Therefore, although discussed with the majority of newly diagnosed patients, HIV partner notification was undertaken with a much smaller proportion of patients. Possible reasons for this were obtained from the medical record review and on discussion with the attending health adviser. Among the 274 patients who had no evidence of partner notification being undertaken, health advisers had considered it inappropriate to undertake partner notification with 63 (23%) patients either because they were physically unwell or because they appeared to be ‘too emotionally distressed’. This was the most common reason given for not undertaking partner notification (Fig. 2). Sixty (22%) patients defaulted or transferred care to other clinics after being diagnosed HIV-positive. This limited the continuity of care and follow-up of partner notification. Nineteen (7%) patients refused to notify any partners.

Fig. 2: . Distribution of reasons for not undertaking HIV partner notification among 274 eligible patients. ^*Other reasons included confidentiality concerns, partner(s) abroad, communication difficulties and manpower limitations. PN, Partner notification; HA, health adviser.

Index patients recruited to the project

Although 197 patients had documentation of HIV partner notification being undertaken, only 70 (36%) patients were recruited to the more detailed evaluation. The remaining 127 were not recruited to the study because of the following: (i) health advisers/physicians considered them to be too emotionally distressed (22%) or too ill (9%); (ii) they defaulted from clinic (19%); (iii) were not seen by a health adviser (16%) or refused to see a health adviser (11%); (iv) the index patient had confidentiality concerns (6%); (v) other reasons (16%). Only 1% of patients specifically refused to participate in the evaluation.

Table 1 shows the demographic profile of the 70 recruited index patients. Index patients were more likely to have been recruited from clinics outside London than within London (64 versus 36%; P < 0.05). There were no significant differences between the recruits and non-recruits with respect to gender and routes of transmission. Amongst index patients, the time spent by health advisers talking about sexual partners and discussing the practicalities of partner notification ranged from 15 min to 10 h (median, 60 min).

Table 1: . Characteristics of HIV-infected index patients.

Time since seroconversion and partnerships

Thirty-eight per cent of index patients thought that they had become HIV-infected within the preceding year, 31% in 1–3 years preceding diagnosis, and 31% thought they were infected over 3 years previously. The median time since possible seroconversion was 2 years (range, 0–13 years), and in this period index patients reported a median of two sexual or drug injecting partners (range, 0–350). The median number of partners considered to be at risk of acquiring HIV infection (through unprotected sexual intercourse or injecting drug use) in the 2 years preceding diagnosis was one (range, 1–60).

At the time of diagnosis, 64% (45 out of 70) of index patients were in a current regular relationship (i.e., index patient had sexual contact with their partner three or more times and intended to see them again). Fifty-five per cent (36 out of 65) had one or no sexual partners within the previous year, and 31% (20 out of 65) had two to five partners during this period. Seventy-two per cent of index patients felt that none or only one partner had been at risk of infection in the year preceding diagnosis.

HIV testing and diagnosis

Sixty-seven per cent (47 out of 70) of index patients had never had a previous HIV test and 49% (34 out of 70) had been tested because of concerns about possible HIV-related symptoms. On initial clinical examination, five (7%) were diagnosed with seroconversion illness, 39 (56%) index patients were asymptomatic, and 26 (37%) were symptomatic for HIV disease. At the time of diagnosis, 70% (12 out of 17) of heterosexual male index patients were symptomatic for HIV compared with 18% (seven out of 38) of homo-/bisexual male index patients (P < 0.05).

Contacts identified for notification

Sixty-five index patients identified 158 contacts for notification (five index patients were unable to name any contacts). Contacts were predominantly male (64%) with a median age of 30.5 years (range, 1–56 years). Unprotected sexual intercourse was the most common (97%) mode of transmission through which contacts may have been exposed to HIV. In total, 71 (45%) contacts were notified, eight by provider referral and 63 by patient referral. The time spent by the health adviser with provider referral ranged from 10 min to 5 h (median, 2 h).

Among the 87 contacts who were not notified, 46 (53%) had insufficient locating information (e.g., address, telephone number), 21 (24%) contacts were already known to be HIV-positive, index patients refused to notify nine (10%) contacts, and three (3%) contacts were already known to be HIV-negative or not at risk. Other reasons given for not notifying seven (8%) contacts included the death of partners and children at risk through vertical transmission. No reason was given for one contact.

Outcomes of notification

Of the 71 notified contacts, 43 were not verified (i.e., information given by index patient only and not confirmed by a health adviser). The outcomes of HIV counselling and testing for these contacts were obtained from their index patients (Fig. 1). However, detailed information was obtained from 28 contacts who were seen in participating clinics. They were more likely to be regular partners of an index patient than those for whom notification was not verified (93 versus 53%). They were also more likely to be heterosexual than homo-/bisexual (64 versus 36%). There were no other significant differences between verified and unverified contacts with respect to their gender, race, residence or risk/exposure categories.

Twenty-five (89%) of the 28 verified contacts accepted pre-test counselling and 27 contacts were tested for HIV infection. Five contacts were found to be HIV-positive, only one of whom considered himself to be at high risk for acquiring HIV. The demographic and risk characteristics of the five HIV-positive contacts are summarized in Table 2. All were notified through patient referral and were either current or past (within last year) regular sexual partners of an index patient.

Table 2: . Characteristics of the HIV-positive contacts identified through partner notification.

Discussion

The findings of this national evaluation indicate that although health-care workers in GUM clinics had discussed partners with most (75%) newly diagnosed HIV-positive patients, documentation of the outcomes of partner notification was found with only a minority (42%). The study was pragmatic and identified a number of logistical factors that limited the feasibility and effectiveness of partner notification within this setting. Nevertheless, amongst those patients with whom partner notification was undertaken, contacts in the wider community who were at high risk of acquiring HIV infection and who were unaware of their HIV serostatus were identified.

Recruitment to the detailed outcome evaluation was disappointing (14% of eligible patients, 36% of those who had partner notification), despite having trained health-care workers, undertaken regular clinic visits and provided reminders on recruitment. A combination of factors may have contributed to this. First, many HIV-positive patients who were eligible for partner notification were not seen by health advisers. Often, physicians did not routinely refer eligible patients to a health adviser and occasionally patients refused to see a health adviser. In addition, HIV-positive patients who attended clinics irregularly, who defaulted from clinic or who transferred care [¹⁴] often missed being referred to a health adviser and having partner notification raised or completed. Second, many health advisers did not undertake partner notification. Our data suggest that patients being assessed as physically unwell or emotionally distressed were the commonest reasons for not undertaking partner notification (22%) and not recruiting patients to the study (31%). Although patients may have been referred to a health adviser for partner notification, more pressing issues (e.g., coping with the diagnosis, emotional crises) would often be dealt with initially. Partner notification was then deferred to a more ‘appropriate’ time, which risked it not being undertaken at all. Third, health advisers' own attitudes towards partner notification [¹⁵], coupled with their concerns about its acceptability to patients [¹³], may have constrained partner notification and recruitment to the study.

Despite these limiting factors, amongst the 70 recruited patients, we found HIV partner notification to be effective in identifying high-risk contacts and previously undiagnosed HIV infections in this setting. Just under one-half (45%) of the 158 contacts identified as being at increased risk of acquiring HIV were notified, and patient referral was the most common mode employed. The results of the outcomes of notification can be contrasted with other partner notification studies performed in the United Kingdom, the United States and Europe [^7,8,16–19]. In our study, approximately 2.3 partners were named per index patient interviewed. In Sweden, Giesecke et al. [⁷] identified 564 contacts from 365 index patients (ratio, 1.5), and Pavia et al. [¹⁹] in Utah identified 890 contacts from 244 index patients. Findings in other studies have ranged from 1.0 to 5.3 partners named per index patient. We found that the effectiveness of partner notification in finding named contacts was primarily limited by the lack of sufficient locating information (e.g., name, address). As a result, current partners or ex-partners (within the past year) were the contacts most likely to be notified, irrespective of the length of the agreed notification period. When undertaken within this subgroup, partner notification may be easier, and more effective.

As mentioned above, over 96% (27 out of 28) of verified contacts accepted HIV counselling and testing after being informed of their exposure risk. Even though the majority of these contacts were HIV-negative, they nevertheless benefited from having focused counselling with a health-care worker, which has been shown to be effective in improving condom use, adoption of safer sex behaviours and sexual risk reduction [²⁰]. The prevalence of undiagnosed HIV infections among the notified contacts, at about 18% (five out of 28), was comparable with other studies. In the Swedish study [⁷], 15% (53 out of 350) of notified contacts were diagnosed HIV-positive for the first time, whereas Rutherford et al. [⁸] in San Francisco reported a corresponding figure of 21% (seven out of 34 notified contacts).

All five HIV-positive contacts had been regular sexual partners of the index patients, none had been HIV tested previously, and only one thought themselves to be at high risk of acquiring HIV. Partner notification in this setting was therefore successful in identifying persons who possibly would not have been identified through other means (e.g., HIV testing campaigns). Thus, as a targeted intervention, partner notification can be a valuable and effective part of HIV detection and prevention strategies. This may become increasingly important as the benefits of HIV diagnosis for the individual increases with the availability of effective antiretroviral therapy [²¹].

Our study had several limitations. We relied on medical records and communication with health-care workers to determine whether or not HIV partner notification was discussed or undertaken with eligible patients. Under-reporting may have occurred if this was not routinely documented. However, partner notification outcomes for STD are routinely documented in medical records in order to facilitate call-back, follow-up and resolution of contact tracing efforts. We believe that this method of review for HIV partner notification was justified and proved an adequate indicator of HIV partner notification activity.

Geographical differences in attitudes and practice of HIV partner notification in England have been previously documented [^13,15]. These may have influenced recruitment to the evaluation and led to unavoidable selection bias (64% of index patients were recruited from clinics outside London). In addition, because many health advisers recruited patients who were physically able, emotionally stable and willing to participate, this could have further contributed to selection bias.

Furthermore, the study design may have contributed to the poor recruitment to this evaluation. The pragmatic nature of the study meant that health advisers were asked to recruit, obtain informed consent and interview index patients on a potentially sensitive and difficult issue. Many health advisers found these tasks particularly difficult during their consultation with patients, which may have led to the ‘less difficult patients’ being recruited to the study.

Our results may not generalize to settings outside the United Kingdom. However, our findings have important implications for the development of strategies aimed at improving the uptake and delivery of HIV partner notification services. First, a distinction must be made between an informal discussion of partners with patients, and initiating formal partner notification whereby partners are identified, notified and counselled and the process monitored and evaluated. The latter requires a more proactive stance from health-care workers involved in partner notification, who in turn need to be adequately trained and supported. Second, attention must be given to the patient management process needed to facilitate partner notification within GUM clinics. This requires clarification of the roles and responsibilities of health advisers and physicians in partner notification. Doctors are more likely to see patients on a regular basis and could introduce issues concerning partner notification before referring patients to a health adviser. All newly diagnosed HIV-positive patients should be routinely referred to a health adviser to discuss partners and sexual health, whether or not they require any psychosocial support. Health advisers and doctors should be responsible for returning to unresolved partner notification if it is delayed or deferred. Provider or contract referral can be offered at this stage if patient referral is difficult. Third, the development of local HIV partner notification policies must be encouraged. Current national guidance on HIV partner notification [¹²] could be a useful starting point. The use of clinical audit to change and monitor HIV partner notification practice should also be supported.

In conclusion, there are many practical issues facing health-care workers and patients involved in HIV partner notification. However, with new combination therapies, improved patient counselling, and effective behavioural interventions, the value of partner notification in identifying undiagnosed HIV infections is increasing. At present, partner notification is seldom carried out, but appears to be successful when it is. Our evaluation has shown that an active strategy for HIV partner notification was able to identify new HIV-positive patients who were unaware of their status and for the most part, their risk of infection. Further studies are now needed to examine the acceptability of this strategy to patients and health-care workers and to evaluate the impact of notification on subsequent behaviours.

Acknowledgements

We acknowledge with gratitude the help of the many health advisers and GUM consultants who participated in this project: R. Hodgekins, J. Home, S. Drake (Birmingham Heartlands Hospital, Birmingham); P. Browne, D. Palmer, D. Wass, A. Nayagam, S. Tchamouroff (Royal Sussex Hospital, Brighton); W. Cox, S. Dowling, J. Llwelleyn, J. Nelki, J. Rice, P. Horner (Bristol Royal Infirmary, Bristol); J. Hearn, L. Maloney, C. Carne (Addenbrooke's Hospital, Cambridge); E. Groves, G. Jacoby, C. Marks, A. Whitehouse, A. Wade (Coventry and Warwickshire Hospital, Coventry); R. Ali, A. Rayner, P. Robinson, S. Thompson, M. Waugh (Leeds General Infirmary, Leeds); C. Brown, R. Evans, C. Jones, S. Ogilvie, A. Roberts, O. Arya, P. Carey (Royal Liverpool University Hospital, Liverpool); A. Stewart, A. Ash (Ealing Hospital, London); J. Buggy, J. Carmodey, G. Hamer, K. Hodgetts, V. Lavick, M. Milton, H. Nicholas, A. Price, P. Young, T. McManus (King's College Hospital, London); R. Barrett, A. Billington, E. Callow, S. Chippendale, L. French, A. Large, S. Paragreen, H. Wilson, E. Allason-Jones, D. Mercey, I. Williams (Middlesex Hospital, London); T. McGrath, C. Simpson, S. Swain, P. Simmons (St Bartholomew's Hospital, London); P. Brewer, C. Johnson, J. Railton, W. Majewska, M. Read, E. Davidson (St George's Hospital, London); S. Feasdale, C. Harman, L. Malloy, N. Strahan, D. Barlow (St Thomas' Hospital, London); Y. Kirlew, J. Quilligan, R. Seery, P. Tovey, R. Hillman, J. Sweeney (Royal London Hospital, London); J. Sharples, B. Skirrow, M. Bhattacharya (Manchester Royal Infirmary, Manchester); J. Atley, C. Faldon, E. Gould, S. Walkinshaw, R. Pattman, K. Sankar, P. Watson (Newcastle General Hospital, Newcastle); R. Chowdhury, L. Glencorse, M. Kersey, K. Moss, R. Shaw, A. Edwards (The Radcliffe Infirmary, Oxford); S. Mulhall, J. Scott, R. Wallace, G. Morrison (Freedom Fields Hospital, Plymouth); G. Bell, T. Cox, M. Horton, S. Peters, R. Poll, G. Kinghorn (Royal Hallamshire Hospital, Sheffield). Advisory Group Members: Dr S. Drake, Ms S. Kent, Dr V. King, Ms W. Majewska, Ms D. Palmer, Mr N. Partridge, Dr P. Simmons, Dr M. Waugh. Special thanks to the Department of Sexually Transmitted Diseases, Division of Pathology and Infectious Diseases, UCL Medical School and Camden and Islington Community Health Services NHS Trust for the support provided.