Article Text

Original article
Is Accelerated Partner Therapy (APT) a cost-effective alternative to routine patient referral partner notification in the UK? Preliminary cost–consequence analysis of an exploratory trial
  1. Tracy E Roberts1,
  2. Angelos Tsourapas1,
  3. Lorna Sutcliffe2,
  4. Jackie Cassell3,
  5. Claudia Estcourt2,4
  1. 1Health Economics Unit, School of Health and Population Sciences, University of Birmingham, Birmingham, UK
  2. 2Centre for Infectious Diseases: Sexual Health and HIV, Barts and The London School of Medicine and Dentistry, London, UK
  3. 3Division of Primary Care and Public Health, Brighton and Sussex Medical School, Brighton, UK
  4. 4Barts and the London NHS Trust, London, UK
  1. Correspondence to Professor Tracy E Roberts, Professor of Health Economics, Health Economics Unit, University of Birmingham, Edgbaston, Birmingham B15 2TT, UK; t.e.roberts{at}bham.ac.uk

Abstract

Objectives To undertake a cost–consequence analysis to assess two new models of partner notification (PN), known as Accelerated Partner Therapy (APT Hotline and APT Pharmacy), as compared with routine patient referral PN, for sex partners of people with chlamydia, gonorrhoea and non-gonococcal urethritis.

Methods Comparison of costs and outcomes alongside an exploratory trial involving two genitourinary medicine clinics and six community pharmacies. Index patients selected the PN method (APT Hotline, APT Pharmacy or routine PN) for their partners. Clinics and pharmacies recorded cost and resource use data including duration of consultation and uptake of treatment pack. Cost data were collected prospectively for two out of three interventions, and data were synthesised and compared in terms of effectiveness and costs.

Results Routine PN had the lowest average cost per partner treated (approximately £46) compared with either APT Hotline (approximately £54) or APT Pharmacy (approximately £53) strategies. The cost–consequence analysis revealed that APT strategies were more costly but also more effective at treating partners compared to routine PN.

Conclusion The hotline strategy costs more than both the alternative PN strategies. If we accept that strategies which identify and treat partners the fastest are likely to be the most effective in reducing reinfection and onward transmission, then APT Hotline appears an effective PN strategy by treating the highest number of partners in the shortest duration. Whether the additional benefit is worth the additional cost cannot be determined in this preliminary analysis. These data will be useful for informing development of future randomised controlled trials of APT.

  • Accelerated partner therapy
  • cost–onsequence analysis
  • sexually transmitted infection
  • partner notification
  • exploratory trial
  • chlamydia trachomatis
  • neisseria gonorrhoeae
  • non-gonococcal urethritis
  • economic analysis
  • chlamydia infection
  • reproductive health
  • sociology
  • opportunistic infection
  • heterosexual behaviour
  • gum services
  • risk profiles
  • service development

Statistics from Altmetric.com

Introduction

The major economic consequences and burden of sexually transmitted infections (STI) and related disease are in women of reproductive age. The asymptomatic nature of infections such as Chlamydia trachomatis (chlamydia) mean that treatment is often delayed, leading to an increased risk of complications and transmission to partners. Complications in women include pelvic inflammatory disease, ectopic pregnancy and infertility, along with neonatal complications in their children.1 The annual National Health Service (NHS) costs of untreated chlamydial infection are around £100 million.1 Less specific information is known about other similarly treatable STIs.

Partner notification (PN) aims to reduce duration of infectivity by informing the infected person's sexual partners of the exposure, offering diagnosis and treatment, and providing advice about prevention.2 Current PN practice in the UK is outdated,3 and a need for research and development in this field has been recently stated by the National Institute for Clinical Excellence.4 The importance of exploring new PN strategies such as Accelerated Partner Therapy (APT) is based on the premise that the strategies that identify and treat partners the fastest are likely to be the most effective in reducing reinfection and onward transmission. In the USA, 61% of index cases receiving Patient Delivered (expedited) Partner Therapy reported that all of their partners were very likely to have been treated compared to 49% of index cases in the standard referral arm.5 In a review of economic studies evaluating screening for C trachomatis, only three out of 59 published economic evaluations had included any form of PN strategies in their analysis.6

The objective of this study was to obtain cost data for APT strategies to target the most common treatable bacterial STIs evaluated in an exploratory trial and to use these data in a preliminary economic evaluation.7 Since this was an exploratory trial, to maximise the utility of the study, the focus was not restricted to chlamydia alone but included both non-gonococcal urethritis (NGU), which occurs in men only, and gonorrhoea, which presents in men and women. All three STIs have similar presentations. Chlamydia is the STI that has traditionally received most attention because it was perceived as the main cause of NGU before specific diagnostic tests were invented. Here, we report the results of a preliminary economic evaluation which compared two new methods of PN (APT Hotline & APT Pharmacy) with routine PN. The analysis was carried out alongside the exploratory trial and conducted from the perspective of the NHS, and so only direct health service costs are included.8 These results will help to refine APT strategies and develop a protocol for a future randomised controlled trial (RCT) comparing the outcomes of APT with standard PN.

Methods

The exploratory trial is reported in detail elsewhere.7 Following the predetermined criteria, eligible index patients who had given consent to participate in the study were offered three methods of PN: (1) APT Hotline—telephone assessment of their sex partner by a clinic-based nurse-qualified health adviser. In Clinic B, most of the health advisers were not nurse-qualified. This meant that a clinic doctor needed to conduct a short additional telephone consultation with the patient to ensure safe prescribing; (2) APT Pharmacy—assessment of sex partner by a trained community pharmacist; (3) Routine PN (patient referral which included infection specific information, advice that the sex partner should attend the clinic for testing and treatment and, in one clinic, a standard letter detailing antibiotic treatment options for the sex partner to give to his/her general practitioner if appropriate). Each index patient was asked to choose which method they preferred for each contactable sexual partner. Once this had been decided, the nurse-qualified health adviser entered these details on the APT database, which generated a unique relationship APT study number. The index was instructed to provide the partner with all relevant information.7 Once the partner engaged with the allocated PN method, the appropriate healthcare professional explained the study and sought consent from the partner to participate. Any partner who did not like the method of PN chosen for them by the index patient could default to the routine PN but was excluded from the principal economic study analysis.

Partners in the APT Hotline group could either collect the treatment pack from the clinic reception, or the index patient could take the pack to them, which occurred if the sex partner completed his/her telephone assessment before the index patient had left clinic. Partners in the APT Pharmacy group received their treatment packs from the trained community pharmacist at the time of the consultation.

Two genito-urinary medicine (GUM) clinics (Clinics A & B) and six community pharmacies participated in the study. All participating clinics and pharmacies recorded relevant data on the duration of the consultation, the uptake of the APT pack and engagement in the process. The engagement in the APT process was apparent when the partner adhered to the method of PN chosen for them by the index case. In some cases, it was necessary for the index case to be followed up for confirmation about where, if at all, the partner received APT or routine care. All costs incurred as a result of the APT strategies were collected prospectively. Costs from secondary sources have been inflated to current prices using the hospital and community price index. All cost data reported are presented in British pounds in 2008 prices.8

Routine Patient Notification

Resource use associated with routine PN was based on the primary data collected in the Chlamydia Screening Studies (ClaSS) project carried out by one of the research team (TR).6 9 A particular component of ClaSS required collection of primary resource use and some cost data associated with PN in GUM clinics and primary care. For example, the duration of consultation with the health advisor was recorded in ClaSS, to which unit costs of health and social care were applied. Estimates of the resource use and primary costs associated with medication and disposables used in routine PN practice, such as testing kits, were standardised to be comparable with what was being offered by the APT strategies, and thus costs used in the ClaSS project were appropriately adjusted. A detailed explanation of costs associated with routine PN is presented in Web appendix 1.

APT Hotline

The estimated cost of the APT Hotline included the cost of the telephone equipment, the cost of the consultation and the cost of the receptionist for giving out the packs (for those partners who collected the APT pack themselves). In addition to the duration of the consultation, we assumed that the nurse qualified health adviser spent 10 min carrying out administrative work such as filling in forms and passing on relevant information to the receptionist. The cost of the consultation was estimated using the Unit Costs of Health and Social Care 2008.7

APT Pharmacy

The duration of the consultation with the community pharmacist was recorded and the cost per hour used was based on the estimates in the Unit Costs of Health and Social Care 2008.7 In addition to the duration of the consultation, it was assumed, based on direct reports from the study, that, the pharmacist spent 10 min carrying out administrative work such as filling in forms. The APT Pack was collected on the spot, at the end of the consultation with the community pharmacist.

Analysis

Our approach to the analysis represents an illustration of the type of preliminary analysis that would be appropriate if the study had been carried out as an RCT as opposed to an exploratory trial.

The average cost per partner treated for each strategy (as a share of the total cost for the strategy) is estimated as follows—(total cost of strategy/number of partners treated by strategy).

We compared the costs and outcomes associated with all three strategies separately in an analysis that is, typically referred to as a cost–consequence analysis.10 In such an analysis, costs and outcomes are assessed in a disaggregated manner to see if there is any strategy that shows clear dominance. Dominance occurs when one strategy is shown to cost less but is more effective in terms of the outcome achieved compared to another. A strategy is said to be dominated if it costs more but is less effective than the comparator. We explored costs and consequences for two separate cases. The principal focus is presented in Case 1 and this cost–consequence analysis is based on the outcome of whether the index patient's partner(s) was treated by the allocated method of PN.

In Case 2, undertaken for illustrative purposes only, the cost–consequence analysis was extended to include partners treated by any method since treatment was not always achieved using the APT intervention to which the partners had been allocated.7 The information for this was typically obtained through clarification with the index cases and these results are the focus of the exploratory trial.7

Since this is a preliminary economic analysis alongside an exploratory trial, it was not appropriate to carry out sensitivity analysis because the results are illustrative, preliminary and subject to bias.

Results

The cost estimate of a sexual health check-up for the routine arm was estimated at £45.89.

The cost of the APT pack was common to both APT strategies in the current study and differed only in whether the treatment was for chlamydia/NGU or gonorrhoea. The total APT pack cost for chlamydia/NGU and gonorrhoea was £29.48 and £24.31 respectively due to the different costs of the antibiotics used for treating chlamydia/NGU (Azithromycin, £8.95) and gonorrhoea (Cefixime, £3.78).

Full breakdowns of the health service costs of the APT Hotline and APT Pharmacy are presented in tables 1 and 2 respectively.

Table 1

Health service costs of the APT Hotline strategy

Table 2

Health service costs of the APT Pharmacy strategy

The average cost per partner treated by APT Hotline (Clinics A & B combined) is estimated at £54.42 and is presented in table 1. The corresponding average cost per partner treated by APT Pharmacy (Clinics A & B combined) is £53.29 while the average cost per partner treated by routine PN (Clinics A & B combined) is £45.89. Routine PN is shown to have the lowest average cost per partner treated than either APT Hotline or APT Pharmacy. These costs are presented in table 3.

Table 3

Total cost, number of partners treated by allocation method and any method and median time from index diagnosis to partner treated

Table 3 presents the total cost for each APT strategy compared to the total cost for routine PN as an average for the two centres. We also present the number of partners treated by each strategy and for the purpose of illustration only, the number of partners treated by any method. The median time from index diagnosis to partner treatment (in days) for each strategy is also reported. The cost of treatment is part of the total cost for each strategy; therefore, the higher cost for any particular strategy compared to another strategy typically reflects the fact that more patients have been treated.

The principal focus of the cost–consequence analysis (Case 1) is to look at the costs of all three strategies and compare these with the corresponding outcomes in terms of the number and proportion of partners treated by the allocated method. The results presented in table 3 show that although APT Hotline has the highest total cost compared to APT Pharmacy and routine PN, it is the strategy that also achieves the highest number and proportion of partners treated 35% (n=47) compared to the other two strategies, which treated 34% (n=15) and 11% (n=13) respectively. As reported elsewhere, there were no significant associations between choice of PN option and gender, age, ethnicity, sexual behaviour, index infection, relationship type or relationship duration identified in either clinic in the study.7 In both clinics, a proportion of all index patients who reported more than one contactable partner selected different PN options for different partners (25% (9/36) of all index patients in Clinic A, 6% (1/18) in Clinic B).7

The least costly strategy is routine PN, but this strategy has fewest partners treated by the allocated method. The variable component in the cost calculation for each strategy is the test and treatment, so treating more partners costs more. Treating fewer partners will cost less. In terms of the median duration in days, from the point of index diagnosis to partner treatment, routine PN is relatively slow (4 days) compared to both alternative APT strategies (both 1 day). On the basis of these results, there is no strategy that is either clearly dominant or dominated. APT strategies appear more effective in treating partners and at a potentially faster rate compared to routine PN, but this increased effectiveness comes at a price.

In Case 2, for the purpose of illustration only, we present the number of partners treated by any method as these are the numbers presented by the clinical counterpart to this paper.7 When numbers of partners are considered as a proportion of available contactable partners, both APT strategies access more partners than routine PN: 59% and 66% for APT Hotline and APT Pharmacy respectively, compared to 36% for routine PN. As reported elsewhere, treatment was not always achieved using the APT intervention to which the partners had been allocated. Forty-one per cent (33/80) of the treated partners in the APT Hotline group were treated outside the trial (e.g., general practitioner or other sexual health service), as were 48% (14/29) of the treated partners in the APT Pharmacy group and 29/42 (69%) in the routine PN group.7

The absolute number of partners treated by APT Pharmacy (n=29; 66%) is lower compared to both APT Hotline (n=80; 59%) and routine PN (n=42; 36%), although, as a proportion of contactable partners, it is actually the highest. By extending the focus to include all partners treated, even though they may not have received treatment via the strategy to which they were originally allocated, a preliminary economic analysis can introduce bias. If PN is not received via the allocated method, it could only have been achieved through routine PN methods, which could have been within or outside the trial, but the benefit would be attributed to the allocated strategy. However, the extra costs are not included in the costs that are presented because we only know this result through confirmation with the index patient. In an RCT, the inclusion of partners treated by any method—not necessarily that to which the partner was allocated—might be considered in an intention to treat analysis. This highlights the issues that might exist if we attempted to carry out a full economic analysis on an exploratory trial and the problems that would be caused by the absence of randomisation. In an appropriately stratified RCT, one might expect the potential number of contactable partners to be approximately the same across all three groups and any difference to be as a result of the intervention, other things being equal.

Discussion

These preliminary results suggest that both APT Hotline and APT Pharmacy show significant potential for providing a cost-effective alternative to routine PN as currently offered in GUM clinics. This potential shows even though subsequent treatment was not always achieved through the APT pathways.

This is the first UK study to explore strategies and collect data for APT, and its strength lies in the prospective collection of data on costs and outcomes as part of the primary study. The study's important limitation is that the analysis reported is carried out on data collected in an exploratory trial. Thus there was no randomisation of index cases or partners to the alternative strategies, which provides considerable potential for bias in the results, as illustrated by Case 2. Of further interest is why index cases with apparently relatively more contactable partners appeared to prefer the APT Hotline strategy of PN. Such issues need exploration. The clear differences in the approaches to practice were apparent in the difference in costs and outcome achieved by the different centres. Future studies must endeavour to standardise practice across centres to achieve a representative set of results so that there is some possibility of clear generalisability of such results across other communities.

However, our reported approach broadly represents an illustration of the type of first stage analysis that would be appropriate if the study had been carried out as an RCT as opposed to an exploratory trial. It was not appropriate to report preliminary incremental cost-effectiveness ratios based on outcomes such as cost per additional partner treated because of the inherent biases apparent in the alternative strategies, particularly the variation in the number of potential contactable partners in each strategy. The results of the cost–consequence analysis that are presented must be considered illustrative and subject to bias.

Additional limitations arise because in some cases the outcome data relied on reported results from the index patient. Consequently the results are likely to be biased according to both the choices made and in some instances on the success reported by the index patient. Furthermore, whether or not either of these strategies will be adopted by policy makers will depend on whether these costs can translate into greater benefit later on as a result of reduced onward transmission. The impact of these APT interventions on onward transmission can only be fully evaluated in an appropriate economic evaluation based on a transmission dynamic model.6

Low et al8 compared PN in alternative settings, namely practice nurse-led PN compared to routine PN carried out in the GUM clinic. Nurse led PN was found to be more effective at reaching partners and both strategies cost approximately the same. It is not possible to directly compare strategies evaluated in the USA by authors such as Golden et al5 with UK-based strategies because of the different legalities governing the provision of drug therapy and health service contact with potential recipients between countries.

The strategies that identify and treat partners the fastest are likely to be the most effective in reducing reinfection and onward transmission. The number of days when partners are infected avoided as a result of APT strategies can only be estimated if the duration of infection is assumed with some degree of confidence and used appropriately in a model-based economic evaluation.1 APT strategies appear to be both more costly and more effective than routine PN. Whether or not the added value is worth the additional costs needs careful consideration. Future studies need to find ways to achieve absolute confirmation (as opposed to assumption) that all partners engaged in this study receive treatment through their allocated strategy or elsewhere. Since a lower uptake of HIV and syphilis testing was apparent in the APT group,7 future studies will also need to consider the wider impact of any concurrent fall in HIV or syphilis testing as result of new approaches for PN and explicitly consider how such trade offs should be valued by policy makers.

Resource use data collected, costs estimated and the results of this study will be useful for informing development of future RCTs. However, the results reported here are preliminary and policy recommendations will depend on the results of studies that consider the added value of PN strategies, compared to the cost savings, if any, associated with the reduction in onward transmission.

Key messages

  • APT shows strong potential for providing a cost-effective alternative to routine PN as offered in genito-urinary medicine clinics.

  • Standard PN was the lowest cost strategy and APT Hotline was the most costly but also the most effective strategy in treating partners.

  • If the additional costs required to implement APT strategies translate into reduced onward transmission, APT could be an important cost-effective strategy in STI control.

  • APT strategies need further economic evaluation in the context of a randomised controlled trial.

Acknowledgments

We are grateful to the Department of Health, who funded this work through the Sexual Health and HIV Research Strategy Committee of the Medical Research Council (MRC). The opinions expressed in this document are those of the research group and not of the Department of Health or the MRC. We thank the patients, clinicians and pharmacists for participating and promoting the research within very busy clinical services. We would particularly like to thank the following people, who have all helped with development of the clinical study upon which the economic evaluation is based Catherine Mercer, Andrew Copas, Nicola Low, Paddy Horner, Merle Symonds, Michael Clarke, Laura James and Anne Johnson.

References

Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Files in this Data Supplement:

Footnotes

  • See Editorial, p 2

  • Linked article 047258.

  • Competing interests None.

  • Ethics approval Ethics approval was provided by Norfolk Research Ethics Committee, REC: 06/Q0101/3. This relates to the main study. Ethical approval was not required for the health economics work described here.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles