Telephone communication of HIV testing results for improving knowledge of HIV infection status

Abstract

Background

This is one of three Cochrane reviews that examine the role of the telephone in HIV/AIDS services. Both in developed and developing countries there is a large proportion of people who do not know they are infected with HIV. Knowledge of one's own HIV serostatus is necessary to access HIV support, care and treatment and to prevent acquisition or further transmission of HIV. Using telephones instead of face‐to‐face or other means of HIV test results delivery could lead to more people receiving their HIV test results.

Objectives

To assess the effectiveness of telephone use for delivery of HIV test results and post‐test counselling.

To evaluate the effectiveness of delivering HIV test results by telephone, we were interested in whether they can increase the proportion of people who receive their HIV test results and the number of people knowing their HIV status.

Search methods

We searched The Cochrane Central Register of Controlled Trials (CENTRAL), MEDLINE, PubMed Central, PsycINFO, ISI Web of Science, Cumulative Index to Nursing & Allied Health (CINAHL), WHOs The Global Health Library and Current Controlled Trials from 1980 to June 2011. We also searched grey literature sources such as Dissertation Abstracts International,CAB Direct Global Health, OpenSIGLE, The Healthcare Management Information Consortium, Google Scholar, Conference on Retroviruses and Opportunistic Infections, International AIDS Society and AEGIS Education Global Information System, and reference lists of relevant studies for this review.

Selection criteria

Randomised controlled trials (RCTs), quasi‐randomised controlled trials (qRCTs), controlled before and after studies (CBAs), and interrupted time series (ITS) studies comparing the effectiveness of telephone HIV test results notification and post‐test counselling to face‐to‐face or other ways of HIV test result delivery in people regardless of their demographic characteristics and in all settings.

Data collection and analysis

Two reviewers independently searched, screened, assessed study quality and extracted data. A third reviewer resolved any disagreement.

Main results

Out of 14 717 citations, only one study met the inclusion criteria; an RCT conducted on homeless and high‐risk youth between September 1998 and October 1999 in Portland, United States. Participants (n=351) were offered counselling and oral HIV testing and were randomised into face‐to‐face (n=187 participants) and telephone (n=167) notification groups. The telephone notification group had the option of receiving HIV test results either by telephone or face‐to‐face. Overall, only 48% (n=168) of participants received their HIV test results and post‐test counselling. Significantly more participants received their HIV test results in the telephone notification group compared to the face‐to‐face notification group; 58% (n=106) vs. 37% (n=62) (p < 0.001). In the telephone notification group, the majority of participants who received their HIV test results did so by telephone (88%, n=93). The study could not offer information about the effectiveness of telephone HIV test notification with HIV‐positive participants because only two youth tested positive and both were assigned to the face‐to‐face notification group. The study had a high risk of bias.

Authors' conclusions

We found only one eligible study. Although this study showed the use of the telephone for HIV test results notification was more effective than face‐to‐face delivery, it had a high‐risk of bias. The study was conducted about 13 years ago in a high‐income country, on a high‐risk population, with low HIV prevalence, and the applicability of its results to other settings and contexts is unclear. The study did not provide information about telephone HIV test results notification of HIV positive people since none of the intervention group participants were HIV positive. We found no information about the acceptability of the intervention to patients’ and providers’, its economic outcomes or potential adverse effects. There is a need for robust evidence from various settings on the effectiveness of telephone use for HIV test results notification.

Plain language summary

The use of telephone for the delivery of HIV testing results

A large proportion of people do not know they are infected with HIV. Knowledge of one’s own HIV serostatus is necessary to access HIV support, care and treatment and to prevent acquisition or further transmission of HIV. Patients often need to return to the testing site to receive HIV test results and post‐test counselling one to two weeks later after being tested. Frequently, people do not return for their HIV test results, particularly in developing countries. In this setting, barriers such as lack of money, transportation or stigma attached to HIV positive serostatus prevent people from collecting their HIV test results. However, the HIV test results could also be delivered by a single phone call, either by a fixed line or a mobile phone. Given the recent rise in mobile phone use in both developed and developing countries, telephone HIV test result notification could be an effective and feasible method for increasing the number of people receiving HIV test results. The aim of this review was to assess effectiveness of the telephone for HIV test result delivery, compared with face‐to‐face or other methods of HIV test result notification. After a comprehensive search of various scientific databases and other resources, we found only one relevant study. This study was performed in 1998‐1999 in the United States on high‐risk and homeless youth. The participants were offered an HIV test and told that their HIV test results would be available in two weeks. They were then divided into two groups; one that had to return to the testing site to get their HIV test results, and another that had the option of receiving HIV test results either by telephone or face‐to‐face at the testing site. Overall, less than half of participants received their HIV test results. Most participants in the telephone notification group opted for telephone rather than in person delivery of HIV test results.The proportion of youth receiving their HIV test results in the telephone group was significantly higher compared to the face‐to‐face group. However, since none of the participants in the telephone group were HIV positive, the study could not provide information about the effectiveness of telephone HIV test result delivery in people with HIV. In addition, we could not find any information about other relevant outcomes such as participants’ and providers’ satisfaction with the telephone HIV test results delivery, cost or potential harmful effects of this intervention. We urgently need more studies conducted in various settings comparing the effectiveness of telephone to other ways of HIV test result delivery and providing other relevant information in addition to the proportion of people receiving their HIV test results.

Background

This review is one of three Cochrane reviews that examine the role of the telephone in HIV/AIDS services. The other two reviews are titled 'Telephone delivered interventions for reducing morbidity and mortality in people with HIV infection' (Gentry 2012) and 'Telephone‐delivered interventions for preventing HIV infection in HIV‐negative persons' (van Velthoven 2012b). The titles emerged from a preliminary overview of the literature (van Velthoven 2012a). The reviews will provide a comprehensive synthesis of the evidence of the use of telephone consultations in HIV/AIDS services. The timeliness of this review is apparent given the increasing use of mobile phones in all settings and vast opportunities they offer for health promotion and disease prevention (Noar 2012).

Description of the condition

Both in developed and developing countries there is a large proportion of people who do not know they are infected with HIV. For example in the United Kingdom (UK) approximately 26% of the people living with HIV are unaware that they are infected (HPA 2010). In low‐ and middle‐income countries the proportion of HIV infected people reported having an HIV test and receiving the result rose from 17% in 2003‐2006 to 39% in 2007‐2009. However, this means that still 61% of those living with HIV are unaware of being infected (WHO 2010). Organisations have addressed this by the recommendation for routine opt‐out HIV screening for all patients (CDC 2006) and the need to increase efforts for HIV testing (WHO 2010).

Knowledge of one’s own HIV serostatus is necessary to access HIV support, care and treatment and to prevent acquisition or further transmission of HIV. Late diagnosis of HIV is related to increased morbidity and mortality. Testing and counselling helps communities to reduce stigma by normalizing the fact that people know their HIV status and by increasing HIV knowledge through post‐test counselling. Also, knowing how many people are infected in an area assists in estimating how much health care support is needed (WHO 2010).

Patients frequently need to return to the testing site to receive HIV test results and post‐test counselling one to two weeks later after being tested. However, often this does not occur. For example, up to 50% of people tested at clinics in the US do not go back to receive their HIV test results. When a return visit is needed, some people may even be less likely to accept testing (Sullivan 2004). People in developing countries are struggling with the many barriers faced in trying to determine their serostatus, such as transportation problems, a lack of money for testing and stigma, making it less likely for them to return to clinics (WHO 2010).

Description of the intervention

Instead of requiring patients to return to the testing site for HIV test results and post‐test counselling, this can be delivered via a phone call. Either the telephone or a mobile phone can be used for this purpose. A study carried out in Australia reports that policy is failing to achieve high levels of results received and that giving results via telephone may be more effective (Healey 2010). The use of the telephone for providing HIV test results and post‐test counselling has been found to be effective in improving the number people receiving results compared with face‐to‐face notification. A before‐after study evaluated 5517 patients before the option of telephone notification was offered and 3886 patients after the intervention was introduced. People in the group who were offered and accepted telephone results were 2.5 times more likely to get HIV results compared with persons in the group not offered the option to receive results via telephone (McKinstry 2007).

How the intervention might work

Using telephones for HIV testing compared to face‐to‐face delivery of test results can improve access, reduce costs, save time, lessen efforts and reach more people, thereby increasing the effectiveness and cost‐effectiveness. People will not have to travel to clinics which saves them time and money. This is especially valuable for people living in rural areas and those whose health or social circumstances make visits to health care providers difficult (Foster 1999; Lattimer 1998). It offers potential for health care workers to consult with more patients within a given time as telephone consultation can be shorter than a face‐to‐face consultation (Bunn 2005).

The burden of HIV/AIDS is the highest in resource‐limited settings, with often poor infrastructure and transport. Also people living in these settings face additional barriers such as limited financial means for travelling, fear of stigma and partner violence. However, there is a potential for delivering HIV test results by mobile phone as mobile phone uptake is the highest in developing countries. The percentage of mobile subscriptions rose from 53% in 2005 to an estimated 73% by the end of 2010 (ITU‐D 2010). Mobile phones have potential to improve access to care, strengthen communication and may help to overcome barriers such as confidentiality issues (Lester 2006). Organisations such as the WHO have included the upscale of telecommunications technology in resource‐limited healthcare settings as one of their priorities (WHA 2005).

However, the effective and safe use of telephone support in healthcare is reliant on qualified healthcare professionals delivering the intervention to avoid adverse events occurring, such as increased out‐of hours contact and A&E attendance (Bunn 2005). The range of possible benefits that telephones offer in medical encounters and their role in medical practice remains controversial for many clinicians (Car 2003; Sokol 2006; Pinnock 2007). There are some possible risks and downsides to using mobile phones for communicating results of HIV testing including unavailability to monitor non‐verbal cues, limited privacy in certain situations, depletion of battery during a consultation and running out of prepaid credit (depending on who is paying).

Why it is important to do this review

The evidence on the feasibility and effectiveness of telephone notification and counselling of HIV test results is non‐conclusive. Most studies comparing face to face consulting with telephone consultation have used purely observational methods (Bunn 2005). However, these methods suffer from difficulty in eliminating systematic bias. Little is known about the content of telephone consultations used for HIV testing or the quality of the counselling given by telephone in comparison with face‐to‐face consulting (Car 2004). Safe and effective telephone consultations require good training of health care personnel (Higgs 2003). The key approaches for clinicians have to be defined and the skills required of them for efficient practice also need to be ascertained. Considering the potential that the telephone can offer to improve HIV testing, it is important to assess the effectiveness. This review intends to summarise evidence on how the telephone can be used for HIV/AIDS testing to improve knowledge of HIV infection status. We aim to contribute to an evidence base for the effective and safe use of this intervention.

Objectives

Primary objective

To assess the effectiveness of using the telephone for delivering HIV test results and post‐test counselling in all settings and in a population regardless of any demographic characteristics.

Secondary objective

To identify factors of applicability of delivering HIV test results by telephone. We will focus our discussion on the safe use, barriers and facilitators for this intervention (access, costs, ease of use, confidentiality).

Methods

Criteria for considering studies for this review

Types of studies

Randomised controlled trials (RCTs), quasi‐randomised controlled trials (qRCT), controlled before and after studies (CBAs) with at least two intervention and two control groups, and interrupted time series (ITS) studies with at least three time points before and after the intervention. We considered qRCTs and CBAs, because our preliminary search showed that there is a limited number of RCTs. The other study types can provide additional evidence. Including ITS studies is valuable in assessing the ongoing qualities of a new technology requiring an introduction period. We also planned to evaluate relevant trials with economic evaluations.

Types of participants

People who receive HIV test results and post‐test counselling by telephone regardless of any demographic characteristic (e.g. age, gender, education, marital status, employment status, or income) in all settings (e.g. primary care settings, outpatient settings and hospital settings or low‐, middle‐ and high‐income countries).

Types of interventions

We considered studies using a telephone for delivering the intervention. The comparison group was face‐to‐face consultation/usual care (e.g. require patients to return for receiving test results) and other ways of receiving HIV test results ( e.g. by post). Outcome measures for patients who received their test results by telephone were compared with the outcomes of patients assigned to the usual care or control group.

Studies using mobile phones for calling were included. Studies using mobile phone messaging (SMS, MMS) were excluded as this is covered by another Cochrane review (Gurol‐Urganci 2008). Studies which use telephones only as a part of their intervention were excluded when the telephone component could not be separately evaluated.

Types of outcome measures

We included studies if they reported on at least one of the following primary or secondary outcomes.

Primary outcomes

To evaluate the effectiveness of delivering HIV test results by telephone, we were interested in whether they can increase the proportion of people who receive their HIV test results and the number of people knowing their HIV status.

1.  Health care uptake related outcomes (e.g. proportion of people receiving their HIV test result)

2.  Knowledge about HIV status (e.g. proportion of people knowing their HIV status)

Secondary outcomes

To evaluate whether the telephone is an appropriate mode of communication, and factors influencing adoption of this intervention, we looked at the following outcomes:

1.    Patient evaluation outcomes (e.g. acceptability of service, patient‐reported communication with provider, ease and timeliness with which patients are able to obtain and maintain needed services, barriers to services, perceived confidentiality and privacy)

2.    Provider evaluation outcomes (e.g. readiness to use, timeliness and/or convenience, satisfaction, barriers to using the intervention)

3.    Economic outcomes (e.g. cost‐effectiveness, cost‐benefit, resources used)

4.    Adverse outcomes regarding the communication mode (e.g. lower quality of telephone counselling, increased out‐of hours contact for the provider, misinterpretation of information)

Studies that do not meet the above listed inclusion criteria will not be included in this review.

Search methods for identification of studies

We used terms specific to the telephone and HIV/AIDS to create a general search for all three reviews. We consulted a librarian and Cochrane HIV/AIDS Review Group's Trial Search Coordinator for developing the search strategy. Articles in all languages but with an English abstract, regardless of publication year were included. We included both peer‐reviewed studies and grey literature. Two reviewers independently performed the searches. The detailed search strategy for MEDLINE is presented in the appendix.

Electronic searches

• The Cochrane Central Register of Controlled Trials (CENTRAL, The Cochrane Library)

• MEDLINE (Ovid); 1980 ‐ present     

• PubMed Central (PMC)

• EMBASE (Ovid); 1980 ‐ present     

• PsycINFO (Ovid); 1980 ‐ present        

• ISI Web of Science which includes: Social Sciences Citation Index (SSCI), Conference Proceedings Citation Index‐ Science (CPCI‐S), Science Citation Index Expanded (SCI‐EXPANDED), Arts & Humanities Citation Index (A&HCI)       

• Cumulative Index to Nursing & Allied Health (CINAHL); 1981‐present         

• WHOs The Global Health Library (Regional indexes)

Searching other resources

We searched online trial registers for ongoing and recently completed studies (Current Controlled Trials: www.controlled‐trials.com). This database contains the following registers:

‐ Action Medical Research (UK) ‐ subset from ISRCTN Register

‐ Medical Research Council (UK) ‐ subset from ISRCTN Register

‐ NIH ClinicalTrials.gov Register (International) ‐ subset of randomised trial records

‐ NIHR Health Technology Assessment Programme (HTA) (UK) ‐ subset from ISRCTN Register

‐ The Wellcome Trust (UK) ‐ subset from ISRCTN Register

‐ UK trials (UK) ‐ subset from ISRCTN Register, UK trials only.

Reference lists

We searched reference lists of relevant studies being included in the review, relevant systematic reviews and key articles in this area (as described or referred to by experts in the field).

Grey literature

We also searched Dissertation Abstracts International (DAI) and CAB Direct Global Health, OpenSIGLE (System for Information on Grey Literature) and The Healthcare Management Information Consortium (HMIC) database . A preliminary Google Scholar search showed a large number of hits, so we scanned the first 500 results.

Conference proceedings

We searched the following conference databases:

‐ Conference on Retroviruses and Opportunistic Infections (CROI)

‐ International AIDS Society (IAS)

‐ AEGIS (AIDS Education Global Information System)

Data collection and analysis

Selection of studies

We merged search results across databases using Endnote, and removed duplicate records of the same report. Two authors (LTC and SG, MvV and SG) independently examined titles and abstracts according to pre‐specified inclusion criteria to determine study eligibility. Studies, for which eligibility was uncertain on the basis of the title and abstract alone, were examined with the full text. Authors discussed which studies should be finally included. In cases of disagreement a third review author (JC) was consulted.

Data extraction and management

We only found one eligible RCT. We retrieved full text for the included study for which two authors (LTC and MvV) independently retrieved details of the study and intervention characteristics using standardised data extraction forms. Any disagreements were resolved by discussion. The third review author (JC) was consulted in cases of disagreement.

Assessment of risk of bias in included studies

Two authors (LTC and MvV) independently assessed the included trial and performed assessment of risk of bias. Disagreements were resolved by consensus, or with consultation of a third reviewer (JC).

We assessed the bias as described by The Cochrane Handbook (Higgins 2009) and EPOC group (EPOC 2002) for six domains:

1. Sequence generation; whether the study used a mechanism that ensured that the same sorts of subjects received the intervention and control

2. Allocation sequence concealment; whether the study made sure that those who admitted subjects to the intervention and control groups did not know in which intervention the subjects were randomised, before the intervention started

3. Blinding; whether personnel and people assessing outcomes were being kept unaware of the allocation of the intervention after inclusion of subjects (blinding of participants is not feasible in this context)

4. Incomplete outcome data; whether the study addressed attrition, reasons for attrition and how this was handled during analysis

5. Selective outcome reporting; whether all pre‐specified outcomes, described in the methods or the outcomes one would expect, were being reported in the results

6. Other potential biases; relevant for this review were comparability of intervention and control group characteristics at baseline, imbalance of outcome measures at baseline and protection against contamination

Data synthesis

We performed a narrative summary of outcomes presented in the included study.

Results

Description of studies

Results of the search

The joint search strategy for the three Cochrane reviews on use of telephone consultation for HIV yielded 14 717 citations. We did not find any relevant conference proceedings or on‐going trials. We found only one relevant study focused on the use of telephone for HIV test result notification.

We present the search process in form of a flow diagram (Figure 1).

1.

Study flow diagram.

Included studies

We included one eligible RCT in this review (Tsu 2002) (See Characteristics of included studies). The participants in the included study were homeless youth and youth at high‐risk for contracting HIV infection. The study was conducted between September 1998 and October 1999 in the Portland metropolitan area, Oregon, United States. The prevalence of HIV in this population is low, at less than 1%.

There were 351 participants in the included study. They were recruited from the community by peer‐leaders and offered confidential and tailored counselling and HIV testing. Testing sites consisted of existing venues and new sites specifically created for this project. HIV testing was performed by a trained counsellor with OraSure oral HIV test (99% sensitive and 99% specific for the detection of HIV 1). Participants were informed that their test results would be available in two weeks. They were randomised into a face‐to‐face (n=187) or telephone (n=167) HIV test results notification group. The face‐to‐face notification group had to collect their results from the HIV Outside Inside Community Clinic. The telephone notification group had the option of receive HIV test results either by telephone or face‐to‐face. When seeking their results, youth were asked for their initials, date of birth, and unique number, regardless of notification method. Each participant provided verbal and written consent and also information about demographics, risk behaviours, and HIV testing history in an interview using a standardized Oregon Health Division serology form. Indeterminate or positive OraSure test results were to be confirmed by a serum HIV test. Youth who failed to collect their results within six weeks were contacted and referred to an established county health department program experienced in contacting HIV positive persons. The outcome of interest was the proportion of participants who were notified about their HIV test results and received post‐test counselling.

Excluded studies

One study was excluded at the full‐text screening stage (Frank 1997). This study was excluded because it assessed the test performance of home HIV testing compared to conventional HIV tests with venous blood and not the effectiveness of using the telephone for delivering HIV test results and post‐test counselling.

Risk of bias in included studies

Overall, we judged the included RCT to have a high risk of bias. Figure 2 and Figure 3 show a risk of bias graph and summary.

2.

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.

3.

Risk of bias summary: review authors' judgements about each risk of bias item for each included study.

The included RCT was judged to have a high risk of selection bias due to the non‐random sequence generation process, which consisted of assigning participants who received a testing packet with an odd serology number to the face‐to‐face notification and those with an even serology number to the telephone notification group. Blinding of the participants and personnel in the study was not possible. We believe this did not influence outcome and therefore the risk of performance bias was judged to be low. The risk of detection bias was judged to be unclear since we could not find any information in the study about the blinding of the outcome assessors. Furthermore, the study was judged to have unclear risk of attrition bias and risk of reporting bias since the protocol of the study was unavailable to us. We also judged the study to have high risk of bias due to baseline imbalance since there were significantly more female participants in the telephone notification group compared to the face‐to‐face notification group.

Effects of interventions

The study examined the proportion of participants receiving HIV test results and post‐test counselling either face‐to‐face or by phone. It did not provide information about the relevant primary/secondary outcomes from this review such as participants’ and providers’ evaluation outcomes, economic or adverse outcomes.

Overall, 48% of participants were notified about their HIV test results and received post‐test counselling. Significantly more participants received their results in the telephone HIV test result notification group: 58% compared to 37% from the face‐to‐face notification group (p < 0.001). In the telephone notification group, the majority of participants who received their HIV test results did so by telephone (88%).

The length of time between the testing and results notification ranged from 5 to 79 days, with a median of 17 days. The mean length of time between testing and receiving results varied significantly (p < 0.01) between the face‐to‐face notification group (24.11 days) and the telephone option notification group (18.63 days).

Significantly more participants received their HIV test results in the group with ages between 19 and 24 years compared to those between the ages of 13 and 18; in the group who identified themselves as white ethnicity compared to those who identified themselves as non‐white; in the group who said that they used alcohol and/or marijuana were more likely to receive their test result than youth reporting use of “harder” drugs with sexual activity; in the group who reported participating in high‐risk behaviours, and in the group with men reporting having had sex with men. There were no significant difference in the proportion of participants receiving test results in relation to gender, site of the outreach testing event, previous HIV testing, drug use, drug use in conjunction with sexual activity and rating their risk of HIV infection as “high”.

The study could not offer information about the effectiveness of telephone HIV test notification with HIV‐positive participants because only two youth tested positive and both were assigned to the face‐to‐face notification group. Neither of them collected their HIV test results and they were referred to the county health department program.

Discussion

Summary of main results

The only included study, which had a high‐risk of bias, showed that the telephone was an effective method for delivering HIV test results and post‐test counselling compared to the face‐to‐face method in the study context. However, no information was provided about the patients’ and providers’ satisfaction with the intervention, its cost‐effectiveness or potential adverse outcomes related to it.

Overall completeness and applicability of evidence

We performed a thorough search of an extensive range of databases and considered as eligible all experimental study designs accepted by the Cochrane Collaboration. Nevertheless, we only found one eligible trial that assessed the effectiveness of telephone delivery of HIV test results compared to face‐to‐face test results notification.

The study was performed in the United States and included high‐risk and homeless youth, but in a setting with low HIV prevalence. Although the included study showed that telephone was more effective than in‐person test result notification, it was focused on a specific group of participants from an upper‐income country. Therefore, the applicability of these results on other participants and settings, particularly low‐income countries is unclear. Furthermore, the two participants that had a positive HIV test result were both assigned to the face‐to‐face notification group. Therefore, the study cannot provide information about the effectiveness of telephone as a HIV test result notification method in HIV positive people.

The included study did not provide any information about the patients’ and providers’ acceptability of telephone as a method of HIV test result delivery, its cost‐effectiveness or potential adverse outcomes.

Recently, a rapid rise of mobile phone use has been observed in developed as well as developing countries (ITU‐D 2010). At least theoretically, this provides a great opportunity for improvement of health and healthcare in resource‐limited settings (Lester 2006). Given the high prevalence and poor treatment outcomes of HIV infection in developing parts of the world, this technology could prove to be highly beneficial in diagnosis, prevention and treatment of HIV. However, the existing evidence does not offer sufficient information on the effects of telephone use for HIV test results notification and counselling either in developed or developing countries.

Quality of the evidence

The only included study in this review with 351 participants had a high risk of bias due to inadequate random sequence generation and baseline imbalance between the intervention and control group. The available evidence is judged to be weak.

Potential biases in the review process

This review has a number of strengths. We rigorously followed the Cochrane methodology, performed a comprehensive search and included all Cochrane accepted experimental study design. It is unlikely that we missed any other relevant study.

Agreements and disagreements with other studies or reviews

A meta‐analysis published in 2006 examined alternative HIV counselling and testing methods to increase knowledge of HIV status (Hutchinson 2006). The authors of this review searched English language databases for studies that used alternative methods for HIV counselling and testing, reported the receipt of HIV test results and had a comparison group. Among the eligible interventions was also telephone notification of HIV test results and post‐test counselling. The authors found two eligible studies on the use of telephone for HIV test results notification: the Tsu 2002 study which we also included and a before‐and‐after study ineligible for this review because of study design (Schulter 1996). The authors statistically pooled the findings of these two studies and reported that after rapid testing, telephone post‐test counselling and home testing were the next most effective methods for increasing the receipt of HIV test results (Hutchinson 2006).

We found another before and after study examining the proportion of people receiving HIV test results after the delivery of HIV test results by telephone had been offered. The proportion of persons who received their HIV test results in the study population was initially very high and remained high over time. However, this proportion declined slightly for persons testing HIV negative. The proportion of persons testing HIV positive who learned their results increased from 85% to 94%, although this result was not statistically significant. After adjusting for age, sex, race, HIV status, and risk group, authors reported that people in the group offered and accepting telephone results were 2.5 (95% confidence interval (CI) 1.7‐3.6) times more likely to get HIV results compared to persons in the group not offered telephone results (McKinstry 2007).

Authors' conclusions

Implications for practice.

The one included study in this review showed that telephone notification of HIV result resulted in significantly more patients receiving their HIV test results compared to face‐to‐face notification. However, this study had a high risk of bias, was conducted on a specific, low HIV prevalence population in an upper‐income country. In addition, it did not provide information about any other relevant outcomes such as acceptability, cost‐effectiveness and potential adverse effects. Due to such scarcity of evidence, we cannot recommend the use of telephone for HIV test result notification in practice.

Implications for research.

There is a lack of evidence on the effectiveness of telephone use for HIV test result notification and counselling. The only eligible study we found had a high risk of bias, was performed in a high income country, on high‐risk, but low HIV prevalence population more than 13 years ago. It is questionable if the findings from the included study are applicable to settings with higher HIV prevalence, in people with HIV or in a low‐income country. Furthermore, we found no information about the patients’ and providers’ evaluation of this intervention, its economic outcomes or potential adverse effect. Although it is plausible and supported by the findings of this study that the telephone HIV test result notification increases receipt of HIV test results, current evidence is scarce. There is a need for high‐quality evidence in various settings and contexts focusing not just on the proportion of participants receiving their HIV test results but also on other relevant outcomes such as intervention acceptability, its economic and adverse outcomes. Given the potential effect that the use of the telephone could have on the uptake of HIV testing results, policy makers should ensure urgent prioritisation of rigorous research on this topic.

History

Protocol first published: Issue 6, 2011
 Review first published: Issue 1, 2013

Date	Event	Description
7 December 2012	Amended	modifications according to the peer‐review comments
13 June 2012	Amended	modifications/editing

Appendices

Appendix 1. MEDLINE search strategy

Number	Search terms
#1	phone*
#2	telephon*
#3	telefon*
#4	telephone intervention
#5	call
#6	calls
#7	calling
#8	telemedicine
#9	teleconsult*
#10	telecounsel*
#11	Telephone[MeSH]
#12	Hotlines[MeSH]
#13	Cellular phone[MeSH]
#14	Cellular phone
#15	Mobile phone
#16	Cell phone
#17	#1 OR #2 OR #3 OR #4 OR #5 OR #6 OR #7 OR #8 OR #9 OR #10 OR #11 OR #12 OR #13 OR #14 OR #15 OR #16
#18	HIV Infections[MeSH]
#19	HIV[MeSH]
#20	HIV[tw]
#21	hiv‐1*[tw]
#22	hiv‐2*[tw]
#23	hiv1[tw]
#24	hiv2[tw]
#25	HIV infect*[tw]
#26	human immunodeficiency virus[tw]
#27	human immunedeficiency virus[tw]
#28	human immune‐deficiency virus[tw]
#29	((human immun*) AND(deficiency virus[tw]))
#30	“Acquired Immunodeficiency Syndrome”[MeSH]
#31	acquired immunodeficiency syndrome[tw]
#32	acquired immunedeficiency syndrome[tw]
#33	acquired immuno‐deficiency syndrome[tw]
#34	acquired immune‐deficiency syndrome[tw]
#35	((acquired immun*) AND (deficiency syndrome[tw]))
#36	#18 OR #19 OR #20 OR #21 OR #22 OR #23 OR #24 OR #25 OR #26 OR #27 OR #28 OR #29 OR #30 OR #31 OR #32 OR #33 OR #34 OR #35
#37	# 17 AND # 36

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Tsu 2002.

Methods	RCT conducted between September 1998 and October 1999
Participants	351 participants received HIV testing. The eligible study participants were homeless and high‐risk youth from Portland metropolitan area, Oregon, United States between the age of 13 and 24. Their ethnicity was mostly white (71%), followed by African‐American (9%), Native American (7%), Hispanic (5%), and Asian (1%). Six percent of youth identified their race/ethnicity as “other” and 1% did not report any race/ethnicity. There were similar proportions of female (49%) and male participants. 184 participants were randomised to intervention group and 167 of them to control group. There were significantly more female participants in the interventions group. There were no other baseline imbalances regarding demographic characteristics between the groups.
Interventions	The participants were recruited by peer‐leaders from the community. Confidential and individualized counselling and HIV testing were performed by a trained counsellor. Testing sites were either incorporated into existing events or created specifically for this event and also included mobile health vans. The administered HIV test was OraSure Oral HIV test (99% sensitive and 99% specific for the detection of HIV 1). Each participant had to provide both written and verbal consent and was given a business card with clinic times and phone number, the client’s unique identifying number from the serology form, and a list of community resources. Participants were informed that their HIV test results would be available two weeks from the testing date and were told about the methods of notification available to them. When seeking their results, youth were asked for their initials, date of birth, and unique number, regardless of notification method. Indeterminate or positive OraSure test results were to be confirmed by a serum HIV test. Intervention: Face‐to‐face notification of HIV test results in which participants were required to receive their HIV test results in the Outside In Community Clinic. Control: Telephone notification of HIV test results group in which participants had the option of receiving HIV test results either by telephone or face‐to‐face.
Outcomes	The outcome was the proportion of youth receiving HIV test results and post‐test counselling.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	Quoted: "Youth who had a testing packet with an odd serology number were assigned to the face‐to‐face notification group and were required to receive their results at the Outside In Community Clinic. Youth with an even serology number were assigned to the telephone notification group and had the option of face‐to‐face or telephone notification for HIV test results." Unsuitable random sequence generation.
Allocation concealment (selection bias)	High risk	No mention of allocation concealment in the study.
Incomplete outcome data (attrition bias) All outcomes	Low risk	The blinding of participants and personnel was not feasible. However we believe it did not  affect the results
Selective reporting (reporting bias)	Unclear risk	No information about the blinding of the outcome assessors in the study.
Other bias	Unclear risk	No information about the missing data in the study.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Protocol of the study was unavailable.
Blinding of outcome assessment (detection bias) All outcomes	High risk	High risk of bias due to baseline imbalance since there were significantly more female participants in the telephone notification group compare to face‐to‐face notification group.

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
Frank 1997	The study compared test performance of home‐HIV test and conventional HIV test with venous blood. It did not assess the effectiveness of using the telephone for delivering HIV test results and post‐test counselling.

Differences between protocol and review

There are a number of differences between the protocol and the present review. These differences are mainly due to the fact that only one study met the inclusion criteria, thus they are mostly in the data analysis section. 

1.   Assessment of risk of bias in included studies: we did not find any eligible studies with study designs other than RCT, so using other tools to assess risk of bias (EPOC 1998; EPOC 2009) as previously stated in the protocol, was not necessary.

2.   Data analysis: if we found more than one study meeting our criteria we would have presented a summary of the treatment effect by study design, type of intervention and setting. Moreover, we would have analysed the outcomes using RevMan and intention‐to treat principle of analysis. If sufficient quality had been available, we would summarise data statistically. We planned to calculate dichotomous data as relative risk (RR) with 95% confidence intervals (95% CIs). For continuous outcomes we wanted to calculate the mean difference and the weighted mean difference (WMD) with 95% CI. For ITS studies we would report mean differences in before and after values. We would also report each study design separately. We wanted to calculate overall results based on a fixed model. If heterogeneity had been detected and it had been appropriate to combine the trials, we would have used a random‐effects model.

3.   Dealing the missing data: Missing data (e.g. studies, outcomes, summary data, and individuals) would have been obtained from original investigators. We would not have performed imputation for missing data. Attrition rates, for example drop‐outs, losses to follow up and withdrawals would have also been investigated.

4.   Assessment of heterogeneity: We would have qualitatively examined the degree of heterogeneity between the included studies in terms of: intervention characteristics, type of provider of the intervention, characteristics of subjects and context of study population and outcomes. We would have looked at the extent to which each of these parameters overlap between the studies. When relevant, we would have statistically test for heterogeneity We would have tested heterogeneity by using the I2 statistic with significance set at >50% and the chi‐squared statistic with significance set at P<0.10. Possible sources of heterogeneity would have been evaluated by sensitivity analysis (Higgs 2003) and a random‐effect would have used if heterogeneity was detected.

5.   Reporting bias: we would have used funnel plots and tests for funnel plot asymmetry to assess the presence of reporting bias.

6.   Data synthesis: We would have performed a meta‐analysis if a group of included studies was sufficiently homogenous. We would have separately analysed the results of the different study types and for the different outcomes. If statistical pooling of results had been inappropriate, we would have undertaken a narrative overview of the results (reference). We would have systematically described each included study according to design, setting, participants, control, and outcomes. We would have also provided additional information such as intervention integrity if available. We would have grouped the included studies according to intervention type. We would have presented each included study in a Characteristics of included studies table and we would have provided a risk of bias assessment and quality assessment of included studies.

7.   Subgroup analysis and investigation of heterogeneity: Subgroup analyses would have been performed if one of the primary outcomes demonstrated statistically significant differences between intervention groups. We would have considered subgroup analyses based on demographic outcomes such as race/ethnicity/nationality, age, gender and HIV risk factors.

8.   Sensitivity analyses to evaluate the effect on effect size of the following factors would have been considered for the following factors: lower quality studies, studies with higher/lower risk of bias (as specified above), studies of low statistical power, unpublished studies, studies in languages other than English, research undertaken in various countries, source of funding (industry versus other) and different statistical models (fixed‐ and random‐effects models). 

9.   Assessment of quality of evidence across studies: we would have used the GRADE approach to evaluate the aggregate quality of the evidence (Higgins 2011) for the primary outcome.

Contributions of authors

Josip Car conceived the idea for the review. Michelle van Velthoven, Sarah Gentry and Lorainne Tudor Car performed the searches, scanning of citations, data extraction and risk of bias. Lorainne Tudor Car wrote the review. Josip Car, Michelle van Velthoven and Sarah Gentry provided comments on the review.

Sources of support

Internal sources

• National Institute for Health Research, UK, Not specified.

  The Department of Primary Care and Public Health at Imperial College London is grateful for support from the National Institute for Health Research Collaboration for Leadership in Applied Health Research & Care Scheme, the NIHR Biomedical Research Centre scheme, and the Imperial College London Centre for Patient Safety and Service Quality.

External sources

• No sources of support supplied

Declarations of interest

None to declare.

New