Include: all humans in the context of an outbreak, epidemic, or pandemic of one of the prespecified infectious diseases of interest (full list provided in Multimedia Appendix 3), using WHO definitions for an outbreak, epidemic, or pandemic [16]. We focused on conditions with the potential to reach epidemic or pandemic status, as it was considered that this could help inform the use of CT to prevent an outbreak from escalating.

The inclusion criteria were as follows: interventions that incorporate CT, defined as the process of starting from a confirmed case diagnosed with an infectious disease (index case), identifying who the index case may have come into contact with, and attempting to communicate with the contacts. CT may be a stand-alone intervention or part of a multicomponent intervention that may involve further management of contacts (eg, treatment or quarantine) or other activities to improve infection detection (eg, education and training of health care staff). Interventions should take place in a community setting (including CT that is led by a health care setting but not limited to that health care setting).

Exclusion criteria were as follows: any other form of PHSMs or intervention to reduce transmission that does not incorporate CT; studies in health care settings where the goal was control of nosocomial outbreaks.

Different CT approaches compared with each other or compared with no CT were included.

Other PHSMs as comparators where these were not part of background measures in both arms were excluded.

Studies with at least 1 measure of effectiveness (ie, measures of transmission of the focal disease, health care use, or mortality) or unintended consequences were included, as detailed in Multimedia Appendix 3.

Experimental studies with a control group, including RCTs, quasi-experimental studies, and before-and-after studies, and observational studies with a control group, including cross-sectional and case-control cohorts were included.

Studies without control groups, reviews of any type, crossover study designs, modeling studies, case series, case reports, and qualitative studies were excluded.

We included only peer-reviewed research published in English, with no restriction on date of publication or setting.

Studies of CT in respiratory transmission focused on TB and COVID-19. Table S1 in Multimedia Appendix 7 summarizes the characteristics, key components, and overview of outcomes for these studies. Further details about the characteristics of interventions evaluated within these studies are available in Table S2 in Multimedia Appendix 7, and detailed outcome data are available in Table S3 in Multimedia Appendix 7. More extensive interpretation is available in the full report [10].

Of the 15 trials focusing on TB, 6 were cluster RCTs [25,42,51,65,67,75], 3 RCTs [43,54,55], 4 controlled before-and-after designs [46-49], 1 interrupted time series [50], and 1 observational study [23]. The TB studies were conducted outside high-income countries, except for 1 study conducted in Taiwan [46]. Nearly half of the TB studies were rated as being of weak quality, with only 2 rated as strong [42,65].

The interventions in 7 of the 15 TB studies incorporated PICT methods [27,29,47,50,53,54,58], 4 were HI [51,55,65,79], 3 involved a combination of PICT and HI [46,52,71], and 1 involved PR [55]. The majority of TB interventions were community-based, and most aimed to counteract the disease by encouraging people to attend health care facilities for screening or by facilitating access to home-based screening; 11 of the 15 interventions involved home visits [23,42,47-50,55,65,67,75]. Home visits varied in nature, but tended to involve health care staff [23,55] or recruited community members [46-48,50,67] carrying out symptom screening at the home. Some studies also carried out testing in the visits [55,65]. A total of 3 studies involved community members [46,47] or research fieldworkers [65] following up patients’ households for monitoring, while incentives were offered to contacts to present for screening in 4 studies [23,25,42,43]. Several studies included training components in which health care workers or key community figures were educated to help carry out the intervention (n=8) [23,46-48,50,54,55,67].

A total of 6 studies specifically focused on improving case finding in younger ages [51,52,54,55,59,75]. Three studies referred contacts aged 5 years or younger irrespective of symptom status [51,52,59]. Three focused their intervention on children aged 15 years or younger [75]. Some interventions were multicomponent, with activities such as attempts to increase awareness of TB [50] and screening children at school [49] being examples of additional components.

TB interventions were usually compared with a form of standard care. Often this was a national program, and in all but 1 comparator this involved passive case finding, whereby potential cases self-report for testing; the 2023 study by Hanrahan et al [43] used household CT. In 4 studies, the control or baseline program was described as a form of directly observed treatment, short-course (DOTS) program that was adapted or enhanced to improve CT and case finding [23,25,46,49]. It is likely that baseline programs used in other studies resembled DOTS, but they instead described their comparators as a national standard of care [25,42,47,50,55]. One study incorporated CT and active monitoring vs CT and passive monitoring [75].

TB interventions tended to be longer or have longer follow-ups than COVID-19 and STI interventions. The median and most frequently reported intervention duration (or time until last follow-up) was 12 months (n=5) [23,47-49,54], with a range of 3 months [51] to 5 years [25].

While all TB studies involved treatment for contacts with a pharmaceutical in both the intervention and comparator groups, some interventions focused more heavily on improving this aspect compared with the comparator. In particular, 7 studies used various approaches to encourage treatment access or adherence [25,46-48,50,54,65]. These approaches included integration of indigenous case managers, community-led home visits, improving access to chest radiography for young children, and provision of prophylaxis.

A total of 5 trials evaluating CT in COVID-19 were included [40,41,64,73,76], based in the United Kingdom, Belgium, Portugal, and Cameroon. Two studies were rated as strong quality [64,76], 1 as moderate quality [40], and 2 as weak quality [41,73]. Interventions in 3 studies were categorized as technology-based CT methods [44,45,80], and 2 incorporated PICT [68,77].

The cluster RCT by Tchakounte et al [76] investigated digitized CT using a smartphone app. This intervention involved individuals registering for an app, where upon receiving a positive test, their information was sent to the CT unit, a clinical interview conducted, and automated text messages sent to notify contacts of their exposure to COVID-19. A total of 2 studies reported on the same natural experiment which occurred due to an error in the UK CT system [40,41], when a technical issue caused around 16,000 positive cases to be missed by the national CT system, thereby delaying the tracing and self-isolation of missed cases. Finally, there were 2 cohort studies with concurrent control groups [64,73]. One study evaluated a standard COVID-19 CT program with contact-eliciting interviews and mandatory quarantine [64]. Symptomatic close contacts were then referred for evaluation and testing [64]. The other program tested different durations of tracing windows, with contacts traced if their last interaction with the index case was 2 days before onset or test for the standard window, and 3-7 days in the extended window group [73]. Comparators were standard care protocols including national CT programs [40,41], as well as manual CT [76] and no CT [64,73]. The median intervention period was 1 week, and 3 studies had a 1-week intervention arm [40,41,73]. Intervention durations ranged from 2 days [73] to 23 weeks [76]. Interventions were delivered by local public health authorities [64], government contact tracers [73], and health care workers [76].

Both studies evaluating the impact of CT on disease prevalence found positive associations [40,73]. Fetzer et al [40], reported there was an association between delayed notifications to quarantine, as a result of a fault in the UK Test & Trace program, and increased disease prevalence and mortality per case traced with a 6-14 day delay (multiple outcomes in Table S3 in Multimedia Appendix 7) [40].

Extending the backward CT window from 2 to 3-7 days led to 42% more cases being detected among contacts [40,73]. However, 3 studies found no impact of CT on case detection rates [41,64,76].

Additional outcomes were reported in the 2 studies evaluating the effect of delayed notification of close contacts due to an error with the UK Test & Trace system [40,41]. In the study by Findlater et al [41], they observed that the delay in CT led to primary contacts of index cases being slightly more likely to be admitted to hospital within 14 days (OR 1.1, 95% CI 1.0-1.2), although there was no impact on patient mortality or hospital admission rates in their secondary contacts. Fetzer et al [40] reported higher mortality rates in the delayed CT group (additional 0.007 deaths per capita; SE 0.003; P<.05). No studies reported on the unintended consequences of CT.

A total of 15 evaluations focused on TB [23,25,42,43,46-51,54,55,65,67,75], and 5 on COVID-19 [40,41,64,73,76]. Studies on TB typically evaluated interventions with the broad aim of encouraging the identification, testing, and treatment of contacts via home visits, active follow-up, incentives, and other measures. There was just 1 study that incorporated CT and active monitoring vs CT and passive monitoring [75]. Findings were inconclusive across outcome domains and studies, but the greatest change associated with intervention was observed for reductions in disease incidence or prevalence. However, the nature of the evidence, which was highly heterogeneous in design and quality, means that only tentative conclusions may be drawn. Five studies of varying designs, quality and outcomes, provide limited evidence about the impact of CT interventions on outcomes in the context of COVID-19.

A total of 35 studies focused on CT in sexually transmitted or blood-borne diseases. Table S1 in Multimedia Appendix 8 provides a broad overview of their key characteristics, trial arm components, and findings. Further detail about the characteristics of interventions evaluated within these studies is available in Table S2 in Multimedia Appendix 8, and detailed outcome data are available in Table S3 in Multimedia Appendix 8.

Of the 35 prioritized studies falling into the sexually transmitted route of spread, 13 included chlamydia [21,22,33-38,52,53,56,59,68,69,77], 13 included HIV [24,26-29,31,32,44,45,58,60,62,63], 6 included syphilis [30,39,44,61,70,71], 6 included gonorrhea [34,53,56,68,72,77], 3 included nongonococcal urethritis [35,52,56], 3 included trichomoniasis [39,57,74], 1 included chancroid and lymphogranuloma venereum [39], and 1 did not specify which STIs were included [66]. A total of 7 trials evaluated 2 intervention arms [24,29,34-36,52,70,71], and 2 evaluated 3 intervention arms [30,62].

The duration of the interventions was recorded in 28 studies [22,24,26-31,33-39,45,53,56-58,62,63,66,68-72,74,77]. The duration ranged from as little as 45 minutes for the counseling session in the study by Mathews et al [66] to 13 months, the whole study period that personnel were introduced in the study by England et al [33], but 82% of the interventions were less than 6 months in duration. Most interventions involved a clinic setting (n=26) [22,24,26-30,33,34,37-39,44, 45,52,53,56,57,62,63, 66,68,70-72,74,77] or took place within the community (n=6) [21,57,60,61,69,71]. One study was based in a prison with community follow-up of partners [31]. Two were primary care–based [35,36,59], 1 an unspecified health facility [32], and 1 a public health facility [58]. Interventions were delivered by a range of different health care professionals, and in all but 11 studies [24,26,34,52,57,59,60,63,68,74,77] this was the same for both intervention and control.

As the interventions were broadly similar across all conditions within this group, they have been summarized together.

One study involved household contact investigation where all household members of the index case were identified and tested, not solely sexual partners of the index case [60]. In all remaining studies, the CT method involved partner notification (PN), where the sexual partners of patients diagnosed with an STI were identified and informed of their exposure to an infection and the need to be tested and treated [21,22,24,26-39,44,45,52,53,56-59,61-63,66,68-72,74,77]. One study also included notification of social contacts [26]. The majority of the PN methods were patient referral notification, where the index patient was responsible for informing their partners of their risk of exposure and referring them to services for testing (n=17) [21,22,29,30,34-39,53, 56,57,59,61,69,70,72,77]. Contact or referral slips were used in 11 of these studies [22,24,26,30,39,56,57,59,70,72,77], and 2 studies trialed the use of a web-based notification service [30,53]. Provider referral (where notification was instigated by a health care professional) was used in 9 studies [24,27,32,44,52,62,63,68,71], and 3 studies involved contract referral, whereby the index case was given limited time to notify partners before the provider stepped in [24,26,74]. In 5 studies, index cases had the choice of PN approach [28,31,33,45,58,66]. In 1 study, provider referral was used if the index case refused to notify their partners [39], and in another the index case could choose which method of PN to use (patient-initiated with nurse support or provider-initiated) [61].

In 4 studies, the nature of the partner elicitation interview with the index patient was varied [44,59,72,74]. A total of 12 studies incorporated reminders or active follow-up as part of their interventions [27,29-31,35-38,45,52,62,68,70,72].

A total of 12 studies adapted the testing process to encourage partners to engage with testing, thereby enhancing CT [21,22,27,29,31,34-38,45,62,69,71]. In 7 of these studies, index cases were provided with sampling kits to pass directly to partners [21,22,29,37,38,45,62,69]. In 2 studies urine sampling was evaluated rather than swab sampling [21,22]. Of the 7 home-based testing interventions, 3 studies provided the means to return samples to the laboratory [21,45,69], whereas partners or index cases were responsible for returning tests to a site in 3 studies [22,29,62]. The study by Choko et al [19] had an additional intervention arm incorporating a financial incentive for positive partners that attended clinic [29]. Estcourt et al [34-38] conducted 3 studies that involved the provision of sampling kits and treatment either after partner consultation through a telephone hotline [35-38] or consultation with a pharmacist [34-36]. Sampling kits and treatment were provided to the index case for their partners in the 2022/2024 study [37,38], and this was also an option for some partners in the 2015/2016 study, while other partners collected the tests and treatment from clinic reception or from the pharmacist [34-36]. Luo et al [62] incorporated an additional intervention arm where testing was completed as a couple in a clinic setting [62]. Home or field testing was offered to partners in 4 studies [27,31,60,71].

Anonymity of the index case was a feature of the interventions in 8 studies [26,30-32,45,62,67,68], 5 of these were HIV studies [26,31,32,45,62].

New or specialist personnel were introduced to deliver PN services in 5 studies [33,52,59,60,63,74]. In 1 study this was the introduction of 2 public health officers to assist with CT [33], and in another a practice nurse–led PN strategy was evaluated [59]. In 3 studies disease intervention specialists were used [52,63,74]. The introduction of different staff allowed for some studies (n=3) to evaluate the effectiveness of delivering a field-based PN service [52,60,74].

A total of 5 studies incorporated additional counseling and education for the index case, in part providing extra time and support to plan for PN [28,33,39,66,77]. Index cases in the study by Chiou et al [28] also had phone access to a counselor during working hours and through an app or via email 24 hours/day [28].

Most of the study comparators (n=25) involved patient referral PN methods [21,22,24,26,27,29-31,34-38,45,52, 53,56-58,60,62,68-70,74,77], and of these, 13 used contact or referral slips, cards, or vouchers as a means of facilitating notification [20,21,23,25,28,29,33,51,57,59,67,69,76]. A total of 3 studies involved provider referral comparators [44,61,63], 2 involved contract referral [71,72], 4 included various PN methods [28,32,33,59]. In 1 study, no detail was given for which method of PN the comparator used [66]. Two comparators involved no CT [27,39]. In the study by Cherutich et al [27], the comparator arm was delayed (by 6 weeks) provider-initiated PN (compared with immediate provider PN in the intervention arm), and outcomes were collected prior to this delayed PN occurring. The study by Faxelid et al [39] did not involve PN.

In 30 studies, the comparator was the usual care or standard of care [21,22,24,26,28-39,44,45,52,53,56-60,66,68-70,72,74,77]. In 11 studies, the intervention group received usual care or standard care plus an additional element in the intervention [28,30,34-38,45,53,66,70,74].

Information (written or online) provided to the index case was a feature of the comparator in 6 studies [30,34-38,68,74]. In 9 studies, the comparator involved a counseling session or interview with the index case, which was face-to-face in 8 studies [24,28,31,52,58,66,72,77] and telephone based in 1 study [44]. The comparator group in 9 studies involved some form of follow-up [28,30,31,37,38,45,58,62,68,70]. This predominantly involved follow-up of the index case (n=7) for feedback and experience on partner notification [28,31,37,38,45,62,68,70]. The comparator groups in 2 studies followed up partners directly [30,58].

This review included 35 prioritized studies focusing on CT interventions in STIs, including 13 in HIV. In all but 1 study, interventions were a form of PN, which was either provider- or patient-led. Such interventions typically incorporated the use of referral slips, CT, active follow-up, and self-testing kits.

Despite the sizeable body of evidence, there was considerable heterogeneity across study design, intervention components, and outcomes evaluated, and there were only 5 studies rated as being of “strong” quality. Across all effectiveness outcome domains (case detection, disease incidence or prevalence, treatment rates, and other outcomes), it was most often observed that interventions had no effect on outcomes, though this was not exclusively the case, with several instances of improved outcomes, as depicted in Figure 2. Case detection, the most frequently assessed outcome domain, was typical of the pattern of results, featuring in 21 studies, with 7 interventions associated with greater case detection rates and the rest showing no differences compared with their comparators. There were only 2 instances in which outcomes related to unintended consequences were reported: 1 showing no difference between intervention and comparator for the risk of intimate partner violence, and 1 instance in which the intervention was reportedly associated with an unfavorable outcome—an increased likelihood of abandonment. Overall, there were 19 instances of improved outcomes across the 69 outcomes reported, with 1 negative outcome. Positive findings were more likely to be observed when provider referral was compared with patient referral.

There was little consistency in findings or patterns of intervention components that may be more likely to be associated with successful outcomes. The heterogeneity of the evidence, in terms of design (both at the study and intervention level), quality, and outcomes, precludes robust conclusions. However, the overall picture for CT in sexually transmitted or blood-borne diseases is one in which interventions either do no harm or may improve outcomes.

We included conditions with epidemic or pandemic potential. In the case of COVID-19, all comparisons drew from a pandemic in progress. Evidence relating to TB generally drew from explicitly endemic contexts. However, a challenge in the evidence related to characterization of the “background” disease context; for example, whether STI CT was implemented in the context of a live outbreak. As a general point, it was nearly always unclear from non–COVID-19 studies how background disease surveillance specifically influenced the design of innovations in CT. Most studies resorted to appeals to health care system efficiency or general disease burden to justify why CT was evaluated.

Related to the above, a key challenge in the evidence was identifying how different approaches to CT might generalize to “true” pandemic contexts, or across conditions and routes of transmission. The substantial literature relating to bacterial STIs and HIV focused on PN, which implies a specific, and easily identifiable, form of “contact” that may be less relevant in a pandemic driven by airborne transmission, for example. Similarly, evidence relating to TB focused primarily on household contacts. Another facet of generalizability related to intervention context. Our review was global in nature, meaning we included substantial amounts of evidence from the Global South. A challenge, but also an opportunity, for public health in well-resourced settings will be to identify how innovations successful in low-income or middle-income country contexts might be relevant in the Global North [109]. Finally, generalizability may be impacted by the human capital, and other forms of capital, that CT strategies require. Many CT strategies included a focus on training and deploying community health care workers, disease intervention specialists, or community health nurses, or culturally embedded local volunteers. These may not be available at speed and in the required quantities in a pandemic context.

This review did not focus on contact yield as a key outcome, preferring instead measures of disease prevalence and incidence, transmission, and treatment, alongside a parallel focus on unintended consequences. Contact yield is an important outcome because it reflects engagement of contacts with public health services, but it does not capture meaningful impacts on health and well-being of contacts or on disease control. It is an important challenge to the field to consider why an outcome that is primarily about health system outputs, rather than population health impact, is so paramount. This is also important because anecdotally, even where interventions reported improved contact yield, this did not always translate to improved prevalence, incidence, transmission, or treatment outcomes.

About half of all studies in this review reported on a case detection–specific outcome, with treatment rates among contacts a close second. Treatment rates were particularly well described in evidence from CT for bacterial or parasitic STIs but will be less relevant for infectious diseases where treatments either do not exist or are not widely available. The ultimate test of CT effectiveness is whether it reduces, or even eliminates, longer-term incidence or prevalence of the condition, or whether it reduces morbidity or mortality where this is a risk. These outcomes were comparatively rare, and mortality analyses specifically were linked to TB and COVID-19 evidence. In addition, the link between treatment rates and improved mortality is driven by adherence to treatment regimens. In conditions such as TB that require extended courses of therapy, CT-driven improvements in treatment uptake may not translate into improvements in incidence, prevalence, and disease-specific mortality [110].

An innovative feature of our review was the inclusion of unintended consequences of CT. It is notable that despite an exhaustive search for a wide range of unintended consequences, only 2 studies presented evidence in this respect. These studies evaluated CT in sexual health contexts and focused on intimate partner violence or relationship breakdown, with neither finding a difference in intimate partner violence but one finding an increase in relationship abandonment attributable to PN. While these are important, it is likely that the generalizability of these findings to other disease contexts is limited. A broader focus on the potential unintended consequences of health and social interventions is important to fully account for the health impacts of CT, especially for conditions with airborne transmission. For example, in situations where enhanced enrollment into TB treatment is rolled out because of CT, suboptimal concordance with treatment regimens may amplify the challenge of antimicrobial resistance in communities [111].