Background

Sexually transmitted infections (STIs) continue to rise worldwide and generate important impacts on society [1]. STIs and their sequelae are shown to not only have an impact on the health sector but also the private resources of those affected, their families and other sectors of society [2]. Living with an STI can, for instance, affect an individual’s productivity and participation in the labour market [3,4,5]. It can also have a considerable impact on an individual’s mental health (i.e. stigma, depression), compromising an individual’s overall quality of life [6,7,8]. The wider societal impacts spilling over to other non-health sectors are also referred to as societal [9], multisectoral [10] or intersectoral costs and benefits (or consequences) [11, 12]. This review will use the term intersectoral costs.

Interventions relating to STIs are essential but complex in nature (like many other public health problems). This complexity can, in part, be explained by the aforementioned wide-ranging, intersectoral impacts of public health programmes, and this can create challenges in adequately capturing them in economic evaluations [12, 13]. The study perspective (or viewpoint) adopted in an economic evaluation ultimately determines the costs and benefits included in the analysis. A societal perspective is increasingly being advocated for economic evaluations of public health interventions as it is expected to capture all relevant costs and benefits associated with an intervention both within and beyond health [14, 15]. Depending on the study objective and stakeholder interests, it may be appropriate to assess costs and benefits from other perspectives (i.e. when estimating the financial costs of an intervention to a specific healthcare system or provider).

Even though a societal perspective has been advocated in methodological guidelines for some time [16], it is not always adopted or, if attempted, tends to be incomplete [17,18,19,20,21]. As indicated above, this can be due to the methodological challenges associated with quantifying the intersectoral impacts of public health interventions [22]. However, evidence suggests that the societal costs associated with STI-related interventions can substantially contribute to the total economic cost burden [2, 3]. This implies that excluding intersectoral costs from analyses could severely underestimate the total cost burden and present incomplete economic information to policy/decision-makers, potentially leading to suboptimal decision-making (i.e. inefficient allocation of resources) [18, 19, 23]. Hence, more comprehensive economic evidence is needed that will allow policy/decision-makers to better understand the total cost burden associated with STIs.

Given the importance placed on the societal perspective and the potential for suboptimal policy decisions in the absence of a comprehensive evaluation, a systematic review of published economic evaluations which adopted a societal perspective was undertaken. The review aimed to explore the intersectoral costs considered under a societal perspective in economic evaluations of interventions relating to STIs, and the impact of including intersectoral costs on the overall study results. The specific objectives of this review were to i) identify economic evaluations of interventions relating to STIs which aimed to include a societal perspective; ii) analyse the intersectoral costs (i.e. costs broader than healthcare) included; iii) categorise these costs by sector (i.e. patient/family, productivity, other); and iv) assess the impact of intersectoral costs on the overall study results.

Methods

A protocol for this review was published in PROSPERO, the International Prospective Register of Systematic Reviews database (CRD42019130940). The review systematically followed the Centre for Review and Dissemination (CRD) guidance for undertaking reviews in health care [24] and a five-step approach on how to prepare a systematic review of economic evaluations for informing evidence-based healthcare decisions [25,26,27]. The PRISMA guidelines were followed for the reporting of this review [28] (supplemental file 1).

Search strategy

A search strategy was developed in PubMed for MEDLINE together with an information specialist before adapting for use in other databases (supplemental file 2). Seven electronic databases were searched (1999–2019): MEDLINE (via PubMed), EMBASE (via Ovid), Web of Science (Core Collection), CINAHL, PsycINFO, EconLit and NHS EED. The search strategy was updated for 2020–21 in Medline (PubMed) only. The year 1999 was initially chosen as a starting year to reflect the inception of the United Kingdom (UK) National Institute of Health and Care Excellence (NICE) and their implementation of guidance statements for the conduct of health economic evaluations. Key search terms included terms for economic evaluation and STIs including specific infections.

Inclusion criteria

This review considered both trial- and model-based economic evaluations of any intervention relating to STIs that were conducted in an Organisation for Economic Co-operation and Development (OECD) member country. OECD member countries were chosen due to their similarities in terms of health(care) systems and to better compare the methodology of studies concerned with similar health(care) systems. It focused on full economic evaluations that adopted a societal perspective, including cost-effectiveness analyses (CEA), cost-utility analyses (CUA) and cost-benefit analyses (CBA). The methods of CEA, CUA and CBA have varying theoretical foundations and outcomes are expressed differently. A CEA measures assesses outcomes (effects) in natural units, whereas a CUA assesses outcomes in health utilities such as quality-adjusted life years (QALYs). A CBA assesses both costs and outcomes in monetary values. No restrictions were placed on the type of comparator or outcomes. Participants in an intervention had to be at least 10 years of age to reflect international definitions of the start of adolescence (a period during which individuals establish sexual maturation and sexual activity) [29] (the PICO model is shown in supplemental file 3).

Screening of data and data extraction

Search results were exported into EndNote X9. Citations were de-duplicated following the guidelines by Bramer and colleagues [30]. The study selection was performed by two reviewers (LS, LJ). A systematic process was adopted to guide the screening of studies for inclusion [31]. Stage I: title screening (LS), stage II: abstract screening and categorisation of selected studies by study type and disease group (LS), and stage III: full-text screening (LS). This process was checked by a second reviewer (LJ) and discrepancies were discussed. A standardised data extraction sheet was utilised to record data on study characteristics, intersectoral costs and cost-effectiveness estimates for those studies that adopted a societal perspective in addition to a healthcare and/or provider perspective (to illustrate the difference in study results for each perspective) [27]. Corresponding authors of selected studies were contacted for clarification where it was not clear what types of costs were considered. A PRISMA flowchart illustrates the selection process (Fig. 1).

Fig. 1
figure 1

PRISMA Flowchart

Full size image

Analysis

Data were recorded using Microsoft Excel and Word. Intersectoral costs were identified, extracted and categorised by sector using Drummond’s sector-specific cost classification scheme [16] (Table 1). Drummond and colleagues categorise costs into (i) healthcare, (ii) patient and family, (iii) productivity, and (iv) costs in other sectors such as informal care, educational costs, costs in the criminal justice system, household and/or leisure costs [11]. The reported intersectoral costs were converted to US Dollars and the year 2021, adjusting the values by inflation. This was done using an online inflation tool [32] and a currency converter [33]. A narrative synthesis was performed following CRD guidelines [24].

Table 1 Sector-specific cost classification scheme based on Drummond et al. [16]
Full size table

Results

The search strategy generated 23,895 studies after duplicates were removed. Studies were further limited to those published from 2009 onwards, excluding 6483. Though this was due to a high number of records identified in the databases, the publication date reflects the period of time during which intersectoral costs and benefits have gained more prominence [13, 22, 34]. Titles and then abstracts were screened, and a total of 572 economic evaluations were identified. Studies were further screened to exclude evaluations that considered healthcare or intervention costs only. In total, 48 studies were identified for full text screening and 29 studies were taken forward for data extraction and narrative synthesis (see supplemental file 4 for excluded studies).

Study characteristics

Of the 29 studies identified, the majority focused on HPV (n = 11), HIV (n = 8) and chlamydia (n = 7, of which two focused on both chlamydia and gonorrhoea) (Table 2). The remaining studies were concerned with gonorrhoea (n = 1), hepatitis B (n = 1) and hepatitis C (n = 1). The countries of interest in the selected studies included the United States of America (USA) (n = 10), The Netherlands (n = 8), Canada (n = 3), Sweden (n = 3), Germany (n = 2), Australia (n = 1), Austria (n = 1) and Israel (n = 1). The overwhelming approach adopted was a CEA (considering the study authors’ definition of their study as a CEA and CUA). However, not all studies explicitly stated whether they applied a CEA or CUA approach. Only three out of the 29 studies explicitly reported to have undertaken a CUA [35,36,37] and one study was explicit about having conducted both a CEA and CUA [38]. Modelling was used in all but two studies, which were trial-based [36, 37]. Most of the modelling studies applied a (dynamic) transmission model, which is the preferred method when evaluating infectious diseases [39,40,41,42,43,44,45,46,47]. Twenty-eight studies applied a societal perspective of which 8 additionally adopted a healthcare or payer perspective, or both [37, 39, 40, 44, 47,48,49]. One study used a modified payer perspective [50]. More information is presented in Table 2.

Table 2 Study characteristics
Full size table

The types of interventions varied widely. Vaccination and screening interventions dominated, and involved stand-alone vaccination [45, 50,51,52], vaccination in addition to screening [40, 53,54,55,56], vaccination of girls and boys [43, 47], sex-neutral vaccination [49], screening (stand-alone) [41, 57,58,59], screening in addition to testing [60] and screening after vaccination [61]. Other interventions involved test and treat interventions including PrEP (Pre-Exposure Prophylaxis) [42, 62], testing and linkage to care [37] and opt-out testing strategy [46]; expedited partner treatment [46], population-level treatment expansion [44], (on-demand) PrEP [35, 63], treatment adherence interventions [38], guided internet-based behaviour intervention [36] and financial incentives [39].

Identification, exploration and categorisation of intersectoral costs

Different intersectoral costs were identified, relating to the following sectors: patient & family, informal care, paid labour (productivity), non-paid opportunity costs (productivity), education and consumption (Table 3).

Table 3 Classification of intersectoral costs included in the identified studies
Full size table

Patient & family

Patient and family costs were captured in 14 studies and included patient time, travel expenses, out-of-pocket costs and premature burial costs. Four studies estimated patient time and travel to seek care as part of healthcare costs (i.e. screening, treatment, vaccination) [43, 53, 54, 60]; although it was not clear whether this time was equated to lost productivity. Nine studies included travel costs or expenses paid for by patients/families in their analyses [35, 37, 43, 48, 50, 54, 55, 59, 61]. Most of the studies evaluated travel costs associated with the intervention being evaluated, though this was not made explicit in all studies. Studies were also not always explicit about whether this referred to travel time or financial expenses such as travel fares. Out-of-pocket costs related to costs paid for by patients/families and was accounted for by two studies [39, 56]. Premature burial costs were considered in one study [35].

Informal care

Caregiver support (non-family) and unpaid help by family/friends was captured in four studies [38, 42, 52, 56]. Two focused on informal care costs related to HIV/AIDS care [38, 42], one on caregiver time loss during treatment for cervical cancer patients [52] and one estimated the time taken by family members during patients’ palliative care due to hepatitis B-related cancer [56].

Paid labour (productivity)

Productivity costs in terms of paid labour losses were assessed in 24 studies. The majority of studies measured these in terms of absenteeism (time off work). Of those, one study estimated productivity losses in a sensitivity analysis only [49]. Two of the studies measured productivity in utilities and captured these in quality-adjusted life year (QALY) estimates [42, 56]. Here, productivity was attributable to the HIV-related morbidity and mortality and lost income due to death or disability from hepatitis B-related sequelae, both chronic conditions. Presenteeism was only accounted for by two studies [36, 58]. Few studies estimated lost income [39, 51], lost revenue due to unemployment rate gap [63], fringe benefits [39], early retirement [58], avoided future production loss (in a sensitivity analysis) [62] and intervention-related productivity gains [44, 51].

Non-paid opportunity costs (productivity)

Only four studies explicitly reported capturing the costs associated with non-paid work (i.e. domestic tasks, voluntary work) [36, 38, 39, 52]. It was not clear if/how many studies equated non-paid opportunity costs (i.e. lost leisure time) to lost work hours (labour).

Education

School absence was captured in one study only [38]. It refers to an individual missing out on potential productivity and educational attainment, but no further characteristics were stated for those who missed school. School absence was calculated by adding a unit price (based on the Dutch minimum wage) per hour missed.

Consumption

Future consumption costs unrelated to health were considered by one study [39]. These consumption costs referred to national average age-specific expenditures outside of healthcare and were based on the U.S. Census Consumer Expenditures Survey. The study did not further specify what this entailed.

The impact of intersectoral costs on the study results

All studies that applied a societal perspective in addition to a healthcare and/or payer perspective presented more favourable cost-effectiveness results under the societal perspective (Table 4). Four studies reported that interventions were cost-saving from a societal perspective, whereas they were ‘only’ cost-effective under a healthcare or payer perspective [39, 40, 44, 64]. Two studies found the incremental cost-effectiveness ratios (ICERs) of their interventions decreased when applying a societal perspective in addition to a healthcare or payer perspective [47, 49]. One study found their intervention to be cost-effective from both the health system and societal perspective [52].

Table 4 Comparison of cost-effectiveness results from a healthcare or payer perspective and societal perspective
Full size table

Discussion

This study is the first to systematically review economic evaluations of interventions relating to STIs and explore and categorise the different types of intersectoral costs captured under a societal perspective. It also presents evidence that the inclusion of intersectoral costs has an impact on the overall study results.

Principal findings

This review found that the identified studies took a rather narrow approach to the societal perspective and only considered costs relating to a limited range of non-health sectors. For the majority of studies this meant primarily estimating paid labour losses. For others, this meant the inclusion of patient and family costs in their analyses. Very few studies considered informal care costs and other non-paid opportunity costs. Only one study included educational costs and another captured non-medical consumption costs. These findings indicate that even where a societal perspective is adopted, this may often be limited in scope, potentially omitting relevant costs from other sectors. The theoretical definition of a societal perspective, however, does not limit the potential scope to the aforementioned sectors [65].

Even though the inclusion of intersectoral costs was limited to a few cost sectors, where intersectoral costs were accounted for, this resulted in more favourable cost-effectiveness estimates.

Methodological challenges to the study perspective

A primary reason for studies applying a narrow societal perspective could be the methodological challenges associated with capturing these wider costs such as with data collection processes or unavailable data [13]. The identification, measurement and valuation of intersectoral costs and benefits in economic evaluations is recognised as one of four methodological challenges when assessing public health interventions [66]. This review highlights that despite methodological difficulties, it is important to be transparent and if a narrower societal perspective is applied, this needs to be explained and justified.

Classification of costs

This review’s cost classification scheme was established to assess whether (or not) and to what extent intersectoral costs were considered and reported explicitly and transparently. The findings suggest that there is considerable scope for exploring other wider societal costs in relation to interventions addressing STIs. This would help improve understanding of the wider societal impacts of STI-related interventions and inform the design of future, more comprehensive economic evaluations.

Informal care

This review shows that informal care was rarely captured in the evaluations, but where it was considered, it related to chronic conditions including HIV/AIDS, HPV-related cervical cancer and hepatitis B-related cancer. Where (long-term) care is provided informally this makes the inclusion of such costs in economic evaluations crucial. If informal care is not considered (or discussed as a study limitation) this can omit important information and underestimate the total cost burden. Future research is needed to further investigate informal care costs related to STIs, particularly those that can have chronic impacts (i.e. HIV/AIDS, hepatitis B) and those with severe long-term sequelae (i.e. pelvic inflammatory disease, chronic pelvic pain).

Non-paid opportunity costs

This review also found that the costs associated with unpaid work remained largely excluded from economic evaluations relating to STIs. This was difficult to judge as not all studies were explicit about which cost components/resource items they accounted for when referring to productivity costs – i.e. paid labour, volunteering or household work. A number of study authors were approached to clarify whether productivity losses accounted for paid or unpaid labour, or both. The majority responded that only paid labour losses were included due to missing data or the methodological challenges of including unpaid productivity. These findings suggest that greater transparency is needed when a societal perspective is adopted to clarify which costs and benefits are included/excluded and the justification for these decisions.

Education

School absence was only captured in one study. Absence from school due to an STI or seeking treatment for an STI can refer to potential productivity loss or loss of educational attainment. The study that captured absence from school valued each hour missed at school based on the national minimum wage, as informal care and domestic activities. No other costs relating to education were identified.

Consumption

Where future consumption costs unrelated to health were accounted for it was not clear what this involved. Examples of non-medical consumption costs can include travel expenditures or future costs for housing and food [67]. This adds to the call from this review that more transparency is needed in economic evaluations, in particular on the different cost components included (or not included), to increase consistency in terms of the costs captured and improve comparability of results across studies.

Distinct health impacts

Other potentially important distinct health impacts can include costs in the reproductive health and mental health sphere. However, these impacts were not captured in the selected studies. The prevention of STIs can reduce the risk of cervical cancer, pelvic inflammatory disease and infertility among women [68], which is often related to their sexual, reproductive and psychological health [69]. This implies there might be intangible costs related to STIs and their sequelae such as pain, anxiety and psychological suffering that could have an impact on people’s overall quality of life and contribute to the cost burden. Research shows that intangible costs could potentially outweigh healthcare costs and its inclusion in economic evaluations potentially result in more favourable cost-effectiveness estimates [23]. This study acknowledges there are difficulties associated with measuring and valuing intangible costs and the demonstration of attribution, and more research is needed in this area.

Comparison to other literature

Relatively few economic evaluations related to STIs have adopted a societal perspective. This is in line with recent findings by Bloch and colleagues [70] who assessed how costs and outcomes are measured in economic evaluations relating to interventions to control STIs. Their study revealed that multiple studies did not adopt a broader perspective to account for outcomes beyond health, despite national recommendations advocating to do so [70]. The present review focused on those economic evaluations that did adopt a societal perspective and demonstrated that often this perspective is limited to certain cost sectors, predominantly the labour sector. Kim and colleagues similarly found that the CEAs they considered rarely captured impacts on sectors outside health, but if so, productivity losses were the most commonly estimated [18]. Krol and colleagues’ [71] findings also show that economic evaluations tend to predominantly assess paid labour costs [19]. Unpaid work, in comparison, has tended to receive little attention [71]. The present review confirms that non-paid work is almost entirely ignored, or not explicitly reported, in economic evaluations. As indicated above, sexual health is closely related to other sectors, including education. In 2010, Shepherd and colleagues found that school-based behavioural interventions for the prevention of STIs can improve knowledge and increase self-efficacy [72]. Research by Chong and colleagues [73] showed that online sexual-health education have an impact on an individual’s knowledge and attitudes. Overall, it is evident that public health issues and interventions can impact other sectors of society, and that the application of a societal perspective is important. This has recently been highlighted using COVID-19 as an example and demonstrating the broader societal impacts of such disease on other sectors outside health [74].

Policy implications

A societal perspective is generally recommended to allow for all relevant costs and benefits to be considered and for an economic evaluation to be as comprehensive as possible. However, where economic studies adopt a societal perspective, but this only includes certain costs in certain sectors, relevant societal implications may be ignored. As a result, decisions based on an analysis with a limited scope might not be optimal [18]. As shown in this review, adopting a societal perspective and capturing intersectoral costs relating to STIs resulted in more favourable cost-effectiveness estimates [75]. Again, where diseases such as STIs can be prevented, treated or managed this can have an impact on an individual’s physical, mental and social health and wellbeing, their productivity as well as wider society [75]. In order to improve information communicated to policy/decision-makers all potentially relevant intersectoral costs need to be included in analyses, and if a narrow societal perspective is adopted, the exclusion of relevant costs needs to be made transparent and justified.

Implications for research

The costs considered and included under a societal perspective differed across studies, resulting in heterogeneity of study results. This highlights that there is a need for a clear understanding of which costs were included and excluded under any perspective when reviewing and synthesising the existing literature, or when combining results from different studies undertaken in different settings. This is particularly important because the different elements of costs (i.e. care practices, wages) can differ between countries and time points. When researchers adopt data from the existing literature for use in their own work, they need to carefully assess what costs were captured before the results can be relied on and utilised.

Strengths and limitations

The main strength of this review is that it followed a structured and rigorous process. To our knowledge, this is the first review in sexual health to apply a cost classification scheme in order to explore and categorise the intersectoral costs considered by sector. The classification scheme provides a valuable foundation for the critical appraisal of economic evaluations, in particular with regard to the consideration and identification of societal costs. The results of this study can inform the design of future, more comprehensive economic evaluations of public health interventions, building on the classification scheme presented. Another key strength of this review is the exhaustive search strategy that was developed in cooperation with an information specialist, searching nine databases and a wide range of key search terms relating to sexual health. It is however possible that some relevant search terms may have been missed. In addition, the update of the review focused on Medline only, which may have resulted in some studies being omitted, although this was mitigated by extensive hand searching. A potential weakness of the review is that because of the high volume of studies identified in the databases, an initial screening was undertaken to exclude studies where the abstract suggested that the study adopted a healthcare perspective only. This means that relevant studies could have been missed. Future research could review economic analyses that adopted a healthcare (system) perspective to assess in detail which costs these studies captured, i.e. direct medical costs, patient costs, or other costs. Another limitation of this review is that it focused on OECD member countries. Reviewing studies in non-OECD member countries could have identified other potentially relevant costs associated with interventions relating to STIs. Further, this review focused on STIs that are sexually transmitted and interventions related to infections transmitted other than sexually, could have revealed additional cost sectors.

Future research

Further research is needed to investigate wider intersectoral costs related to STIs that were not (sufficiently) captured in this review but that could be important to inform policy/decision-making. Such research could also help explore the intersectoral costs relating to other sexual health aspects beyond disease such as sexuality, sexual behaviour, and related areas. Given the complexity of sexual health future research could explore wider intersectoral costs relating to STIs that have been considered outside of the health economics literature such as in educational journals or journals relating to social services. Furthermore, future research could explore in greater depth the distinction between intersectoral consequences associated with the disease and the intersectoral costs incurred by the intervention being evaluated, as this was not always fully clear.

Conclusion

This systematic review suggests that economic evaluations of interventions relating to STIs that adopt a societal perspective tend to be limited in scope. This risks omitting potentially relevant intersectoral costs that could be important information for policy/decision-making. There is an urgent need for economic evaluations to be more comprehensive in order to allow policy/decision-makers to make better informed decisions.