Design of the screening programme

STI AIDS Netherlands coordinates the screening programme, and the local Public Health Services carry it out, expanding the screening in a phased implementation (stepped wedge, cluster, randomised design). The Centre for Infectious Disease Control, a section of the National Institute for Public Health and the Environment (RIVM), is responsible for the formal evaluation process. An advisory committee consisting of experts and other relevant parties, including the Ministry of Health, support the executive board of the screening programme. The Ethics Committee of the Free University of Amsterdam (METC number: 2007/239) has approved the study, which conforms to national and international legislation. Participants provide online informed consent. The time frame of the demonstration project is 2008-2010.

Rationale for register-based screening

We chose to organise the screening as register-based (i.e. all individuals registered in the target population are actively approached for participation) as opposed to opportunistic (screening offered during consultations or outreach).

The only two published randomised controlled trials (RCTs) investigating the individual benefits of screening assessed systematic screening [6, 7] and have shown a reduction in pelvic inflammatory disease (PID). A new trial has recently been added to this body of evidence [8]. The programme coverage (or effective test rate) is critical for achieving not only individual benefit (reduction of complications by means of early diagnosis and treatment), but also a reduction of transmission at the population level. The reported population coverage of ongoing opportunistic programmes is limited [9]. In contrast, a Danish study has shown that the eligible population was 11 times as likely to be tested in home-based screening as in the opportunistic testing that doctors offer during routine care [10]. In addition, a recent review shows that the evidence base for opportunistic screening is poor [11, 12]. We have demonstrated the feasibility and acceptability of register-based systematic screening in the Netherlands via the Municipal Health Services in a previous study [3, 13].

Rationale for population selection for screening

The demonstration programme has been implemented in three regions: two large cities (Amsterdam and Rotterdam) and one somewhat more rural area (South Limburg). Each year, all residents aged 16-29 years receive an invitation letter. Our previous research showed that prevalence in the population of highly urbanised areas in the Netherlands is well above 3% [3]. People living in Amsterdam and Rotterdam are advised to participate if they are (or have been) sexually active. In areas of lower population density, the population prevalence is between 0.6% and 2%, so that selective screening based on risk profiles is required to reach a better cost-effectiveness profile [3, 4]. Therefore, eligibility for screening in South Limburg is assessed on the basis of individual risk scores calculated from invitees' information online. The web-based questionnaire asks nine questions about demographics and sexual behaviour, and each item contributes to the risk score (Additional file 1). The invitee can request a test kit only if the risk score reaches a predefined value. The prediction rule for this selective screening is based on previous research [4] and aims at a C. trachomatis positivity yield of 4-5%.

We chose to include both men and women in the screening. Population studies show prevalence rates among men comparable to those among women. Including men in the screening makes them part of the solution rather than the problem [14]. Models show that screening both men and women can have more impact than screening only women [15]. However, the impact of screening both men and women on community prevalence has yet to be demonstrated in real life. In fact, a recent literature review called for more high-quality RCTs to ascertain the impact on individual complications like PID and infertility, as well as the impact on community prevalence [12]. We are expanding our screening area by area, and we are including all eligible residents of a geographically confined area within a limited time. This strategy covers social and sexual networks in order to grasp any community impact of accelerated screening.

Logistics

Figure 1 presents the framework of the logistics. The local Public Health Services send invitations to all 16 to 29-year-olds in Amsterdam, Rotterdam, and South Limburg. Municipal population registers supply the addresses. The order of invitations is cluster-randomised per area or municipality according to postal codes lists within clusters in each of the three participating regions. The invitations contain an information leaflet and a letter, advising those who are (or have been) sexually active to participate. The invitation letters provide a unique ID number (login code) that enables the invitee to request a test kit at http://www.chlamydiatest.nl (Figure 2). The invitation letter and information materials (text and website) have been developed and pretested with the relevant age, sex, and cultural groups. In the lower prevalence area, South Limburg, anyone who visits the website must first do a short online risk assessment, and can only request a test kit if the risk score reaches a predefined level of 6 or more, compatible with an estimated positivity rate of 4-5% (see Additional file 1).

Those who have not reacted to the initial invitation within 4 weeks receive a reminder invitation. A reminder is e-mailed to those who do not send a sample to the laboratory within 2 weeks of their request for a test kit; a second reminder is e-mailed 1 week later. Invitees who send a sample to the laboratory receive an e-mail and/or SMS to inform them when their test result is available online. Chlamydia trachomatis positive participants can print the PDF letter with their test results and information for health care providers. The test results on the website can only be accessed by personal login. If the test results are not accessed within 14 days, a reminder e-mail is sent. Everyone with a positive result who has not checked the test result within 4 weeks is sent a letter with the result at the address provided.

The test kit includes a home sample kit (Figure 3; men: urine sample, women: vaginal swab or urine sample). The participants send the self-collected samples to the regional laboratory in an envelope suitable for mailing biological materials. The three participating laboratories are all certified, take part in quality control programmes and use certified nucleic acid amplification techniques according to the manufacturer's guidelines. The one laboratory that uses pooling retests positive pools individually. The two laboratories that do not pool confirm positive tests with a second test on the same specimen.

The patient's GP provide treatment, but if a participant prefers, the local STI centre provides it. The current sex partners of participants with positive test results are advised to be treated simultaneously or seek prompt testing and treatment. Previous partners are notified predominantly by patient referral. Additionally, participants with positive test results can also use the website, anonymously if they wish, to notify previous sex partners of the last 6 months via e-mail. These partners are offered testing within the existing screening programme. Participants who test positive receive a new test kit automatically after 6 months.

Data collection and evaluation

The programme's central database automatically stores data from all invitees. The database includes data pertaining to progression through the screening process and data from the local population registers, the laboratories, and the company responsible for handling the mailings. The central database enables process monitoring, evaluation, and uploading of regular time-framed overviews of key outcomes and indicators of programme performance. Data obtained from online questionnaires for participants and mailed questionnaires for non-responders, including characteristics such as self-reported sexual history and clinical symptoms, are also linked to the central database to allow for more detailed analyses.

The evaluation of the screening implementation consists of various components. A detailed description of the design of the evaluation is given elsewhere [16]. The first part concerns process evaluation, including non-response and acceptability studies [17]. In the second part, the effects of screening are evaluated. For evaluation purposes, we chose a phased implementation of the screening. This 'stepped-wedge design' enables a sequential roll-out in geographical clusters (neighbourhoods), randomised within a number of periods. The design allows for:

1. Estimation of the effect of one or two screening rounds on the population prevalence of C. trachomatis and self-reported PID.

2. Determination of trends of participation in different screening rounds.

3. Assessment of baseline ecological trends among those who have not received a screening invitation at the end of the second screening round.

Other important components of the evaluation are the modelling of positivity and prevalence, and the cost-effectiveness analyses.

The first screening round

Between April 2008 and February 2009, 261,025 people were sent an initial invitation for the first screening round (140,058 in Amsterdam, 107,806 in Rotterdam, and 13,161 in South Limburg). Overall, 21% of those invited were 16-20 years old, 37% were 20-24 years old, and 42% were 25-29 years old; 52% of the invitees were female. One percent of the invitational letters were returned as undeliverable.

Request rate

The median time lapse between the invitation and request for a test kit was 18 days (range 0-428 days). On average, people requested a test kit 28 days after the invitation [standard deviation (SD) ± 30 days]. Reminders significantly increased the requests for a test kit: 33% of the requests for test kits were received after the reminder was sent (Figure 4).

Effective participation rate

Of those who requested a test kit, 41,637 (79%) sent in a sample. This test uptake of 79% translates into a programme coverage (effective participation or test rate) of 16% of the invited population being tested in the first screening round.

The median time lapse between requesting a test kit and returning a sample to the laboratory was 11 days (range 0-390; mean 17 days, SD ± 19). Reminders significantly contributed to the sending of samples: 39% of samples were received after the reminder invitation was sent. See Op de Coul et al. [17] for more detailed and time-framed monitoring indicators of process performance.

Positivity for Chlamydia trachomatis

The laboratories processed 96% of the samples within the predefined target of 10 working days, and test results were made available online. A total of 1758 people tested positive for C. trachomatis, which resulted in an overall positivity rate of 4.2% (95% CI 4.0-4.4%).

Treatment

Ninety-five percent of the participants checked their test results online, and 90% did so within 7 days after being notified by e-mail that the results were available online. Preliminary analyses of the questionnaires, filled in by 43% of all those who were C. trachomatis positive, indicated that an estimated 91% of them saw a physician (82% went to their GP). Of those who consulted a physician, 98% received treatment. Patient referral is the mode of tracing contacts in the Netherlands, but anonymous, web-based, partner notification via the programme was also possible. The patients with C. trachomatis provided e-mail addresses for notification of 382 non-regular sex partners of the previous 6 months. Altogether, 125 of the notified partners (33%) requested a test kit, of whom 107 (86%) sent in a sample for C. trachomatis testing. Twenty-nine of these samples (27%) tested positive.

Participants who were C. trachomatis positive were automatically sent a new test kit at the address provided 6 months after the initial test. Of these participants, 1152 (68%) sent in a lab sample, and 95 C. trachomatis test results were positive - a positivity rate of 8.2% (95% CI 6.6-9.7%).

Acceptability and usability studies are currently in progress among both participants and non-responders. Preliminary data indicate overall good acceptability, with participants showing positive attitudes towards the screening programme and communication via the website and by SMS. In our postal non-response study, we did not find indications that participation in the screening was hampered by limited access to the internet. Further detailed analyses regarding acceptability, process, and non-response will generate more in-depth insight into factors associated with participation, and coverage. These analyses will be reported separately.