Content » Vol 94, Issue 1

Short communication

Home-sampling as a Tool in the Context of Chlamydia trachomatis Partner Notification: A Randomized Controlled Trial

Lars Falk1,2, Sabina Hegic3, Daniel Wilson3 and Ann-Britt Wiréhn1

1R&D Department of Local Health Care, County of Östergötland, SE-581 85 Linköping, 2Department of Dermatology and Venereology, Linköping University Hospital, Linköping University, Linköping, and 3Local Health Care, County of Östergötland Sweden. E-mail:

Accepted Feb 28, 2013; Epub ahead of print Jun 5, 2013

Chlamydia trachomatis is the most common known bacterial cause of sexually transmitted infection (STI) (1, 2) and an important cause of infertility in women and possibly in men (3, 4). Following a remarkable decrease in reported cases of C. trachomatis, there was a 10–15% annual increase in cases reported to the Swedish Centre of Communicable Disease Control between 1997 and 2005 (5). Increase in the incidence of chlamydia have also been reported in many other countries (6). In Sweden, partner notification of chlamydia-infected individuals is mandatory under legislation passed in 1988 and 2004 (7). One possible reason for this increase in Sweden could be that partner notification may not be fully effective in preventing transmission.

The aim of the present study was to evaluate whether home-sampling could decrease the delay between the time when partner tracing starts (i.e. the meeting between the index patient and a counsellor) and the date of testing (sampling) of sexual partners, compared with conventional testing of partners at a clinic.


This Swedish multicentre study included sexually transmitted disease (STD) clinics in 3 towns (Norrköping, Motala and Västervik), and all C. trachomatis-infected individuals presenting between October 2006 and July 2007 were eligible and were invited to enrol in the study.

The primary index patient, i.e. the first individual diagnosed with C. trachomatis in a new sexual network, was randomized to either a conventional partner notification mode (in which the partners were asked either by the index patient or by the counsellor to attend a clinic for C. trachomatis testing) or to a mode in which a test kit for home self-sampling was posted to them by the counsellor or distributed via the index patient. When sexual partners infected with C. trachomatis became index patients, they were assigned to the same study branch as the primary index patient.

At the STD clinics an informed consent was given to the partner tracer. Written information about the study was sent with the letter requiring the partner to be tested for C. trachomatis at a clinic, according to the law, if the index person was randomized or assigned to the conventional clinic-testing study branch. Current sexual partners who were prescribed antibiotics were excluded from participation.

Sampling of female partners was carried out via combined first-catch urine (FCU) and vaginal samples. Sampling of male partners was via FCU. The sampling date was taken as the end-point.

The Kaplan–Meier approach was used in a 1 minus the survival probability calculation for determination of the time-period from being elicited as a sexual partner until C. trachomatis testing. In an overall analysis the differences between the median times were tested using the log-rank test in the comparisons between conventional clinic-testing and home-sampling. Stratified analyses were carried out for gender and for different sexual partner situations, where the latter was defined in 3 categories: current partner; ≤ 30 days, and > 30 days since sexual contact. Differences in proportions were tested with Pearson’s χ2 test. The significance level was set to 5% for all tests carried out.


Of the 920 index patients eligible for contact tracing, 833 individuals (505 women and 328 men) were eventually enrolled. As the intention was to cluster randomize index patients, approximately half (n =451, 54%) were randomized, i.e. individuals believed at the counselling conversation to be a primary index patient. During the study period 447 sexual networks were revealed, comprising 2,390 individuals. After the initial exclusion, there were 1,693 partners, of whom 1,528 (90.2%) were confirmed to have been tested. Eventually 660 partners were enrolled; 461 men (age range 14–60 years, median age 21 years) and 199 women (age range 14–39 years, median age 20 years). Home self-sampling mode comprised 55 women (14–39 years, median age 19 years) and 160 men (15–60 years, median age 21 years). Conventional clinic-testing tracing mode comprised 144 women (14–38 years, median 20 years) and 301 men (14–49 years, median 22 years) (Fig. S1; available from:

Since cluster-randomization was not possible in practice, many index patients were randomized instead of being assigned to the appropriate study arm. This occurred in almost all sexual networks comprising more than 4 index patients (the range of index patients per sexual network was 1–49, median 2). All calculations are therefore performed at the individual level and not at the cluster level. Analyses of median times to test for C. trachomatis showed a significant difference between conventional mode and home-sampling mode: 15 days in the conventional group and 10 days in the home-sampling group (p < 0.001) (Table I, Fig. 1). The difference was seen in the separate male stratum (conventional clinic-test mode = 16 days, home-sampling = 11 days) and in the female stratum (conventional clinic-test mode = 14 days, home-sampling mode = 7 days). Among persons who had current partners there was no difference in time to test between the 2 test modes (p = 0.903), whereas there were significant differences for those who had had a sexual contact within 30 days (conventional clinic-test mode = 14 days, home-sampling mode = 8 days) (p = 0.038) and those with > 30 days since sexual contact (conventional clinic-test mode = 18 days, home-sampling mode = 11 days) (p < 0.001) (Table I).

Table I. Kaplan-Meier comparison of time from being elicited by the index patient to the counsellor as a sexual partner to C. trachomatis testing, between conventional clinic-testing and home-sampling mode

Total n

Conventional test mode


Comparisons between conventional test mode and home-sampling mode


Median days



Median days









< 0.001















< 0.001

< 0.001

Sexual partner situationa

Current partnera

≤30 days

>30 days




















< 0.001

aMissing values: n = 6.


Fig. 1. Comparison between home-sampling (n = 215) and clinical sampling mode (n = 445) regarding the time from the date a sexual partner was revealed by the index patient at the counselling conversation to the date of testing. The Kaplan-Meier approach was used in a 1 minus the survival probability calculation.



The present study showed that home-sampling reduced the delay to testing of partners compared with conventional testing at a clinic. This is in line with results from a study with self-sampling in a partner notification context in Denmark, where partner notification is not mandatory and no testing is required (8). The benefit of the home-sampling mode was seen when the partners for tracing were not current. The reason that current partners were tested early may be that index patients themselves were involved in notifying the partner, which emphasizes the importance of co-operation between counsellor and index patient, as also reported by Trelle et al. (9). The median time to test was 14 and 16 days, for women and men respectively, tested at a clinic in the present study, which was similar to the results from a retrospective case note audit by Horton on 844 index patients in England (unpublished data provided to Clarke) (10).

Despite the fact that all chlamydia-infected persons in the catchment area were referred to the clinic for partner notification, it was often not possible to determine whether a chlamydia-infected individual was not a primary index patient. The revelation of sexual partners to an index patient is a process, and was often not concluded at the first meeting between the index patient and the counsellor. Thus the study was neither a cluster randomized trial nor a strict randomized controlled study (all index patients were not randomized), since only 46% of index patients were assigned to their cluster. The limited opening hours of the clinics for those partners assigned for conventional clinic testing could favour those assigned to home-sampling and may not reflect the actual readiness for testing.

In conclusion, home-sampling of sexual partners appears to be a successful strategy to significantly reduce the delay in testing cases in which the partner to be tested is not a current partner. Current sexual partners of a chlamydia-infected individual were tested within a short time-period irrespective of the tracing mode.


The authors would like to thank the staff of the STD clinics in Norrköping (especially Gunilla Heed, Margreth Wastesson, Jeanette Groenheit and Annica Andersson), Motala (Evy Adolfsson, Nina Agerhall, Anette Wahlström, Ulrika Olai and Maria Lindgren) and Västervik (Lillemor Gustavsson and Kajsa Lindström). The authors also thank Erik Kihlström, Linköping and Heléna Persson, Kalmar for providing information about the laboratory testing procedure, dealing with home-sampled tests and distributing data to the STD-clinic, and Chris Anderson for linguistic revision and valuable comments.

This study was funded by the Medical Research Council of Southeast Sweden, ALF grants from the County Council of Östergötland.

The regional research ethics committee of Linköping approved the study on 6 September 2006 (M 122-06).

The study was registered in a Swedish worldwide web site for medical research “FoU i Sverige” as document 27331, in 2009 and at number NCT01596946.

The authors declare no conflicts of interest.


1. European Centre for Disease Prevention and Control. Annual Epidemiological Report on Communicable Diseases in Europe 2010, Stockholm: ECDC, 2010.

2. Donovan B. Sexually transmissible infections other than HIV. Lancet 2004; 363: 545–556.

3. Kamwendo F, Forslin L, Bodin L, Danielsson D. Programmes to reduce pelvic inflammatory disease: the Swedish experience. Lancet 1998: 351 Suppl 3: 25–28.

4. Joki-Korpela P, Sahrakorpi N, Halttunen M, Surcel HM, Paavonen J, Tiitinen A. The role of Chlamydia trachomatis infection in male infertility. Fertil Steril 2009; 91: 1448–1450.

5. Swedish Institute for Infectious Disease Control. Stockholm: SMI 2012. Available from:

6. Bender N, Herrmann B, Andersen B, Hocking JS, van Bergen J, Morgan J, et al. Chlamydia infection, pelvic inflammatory disease, ectopic pregnancy and infertility: cross-national study. Sex Transm Inf 2011; 87: 601–608.

7. The Communicable Diseases Act. Stockholm: Swedish Legislature 1988: 1488 and 2004: 168.

8. Østergaard L, Andersen B, Møller JK, Olesen F, Worm AM. Managing partners of people diagnosed with Chlamydia trachomatis: a comparison of two partner testing methods. Sex Transm Infect 2003; 79: 358–361.

9. Trelle S, Shang A, Nartey L, Cassell JA, Low N. Improved effectiveness of partner notification for patients. BMJ 2007; 334: 354.

10. Clarke J. Contact tracing for chlamydia: data on effectiveness. Int J STD AIDS 1998; 9: 187–191.

Supplementary content
Figure S1