METHODS

Setting

Participating BPAS branches included Birmingham Central, Cannock, Coventry, Telford (Midlands) and Chester, Liverpool, Preston, Manchester, and Sheffield (north of England). These locations are representative in size and staffing of those covered by BPAS nationally.

Patient recruitment and specimen collection

Detailed procedures for introducing the study and inviting patients to participate were designed. Patients (private and NHS) who attended for assessment for TOP were offered a test for chlamydial infection. There was no restriction on age and, as it was considered unethical to randomise patients into payment and non-payment groups, all patients were offered the screening test free of charge. The study rationale and procedures were explained to all patients. Verbal informed consent was obtained from all those who agreed to participate. Patients accepting the test were asked to provide a “first catch” urine (FCU) specimen collected after at least 1 hour of urine retention. The sample was collected in a plastic “one ounce” bottle. Patients, whether accepting the test or not, were asked to complete (alone or assisted by a healthcare worker) a questionnaire incorporated into the laboratory request form. This obtained demographic, clinical, and behavioural information and included a means, agreeable to the patient, of communicating the test result. Where a patient declined a urine test, the reason (outright refusal or a previous test) was recorded if given. A question asked respondents whether they would still have had the test if it cost £20 (representing laboratory charges and BPAS administration and patient management costs).

Specimen transport and testing

Urine samples were stored at 4–8°C before transportation by local courier to the nearest PHLS laboratory then, if necessary, forwarded overnight by PHLS transport at ambient temperature to the testing site (Liverpool PHL or Telford PHL). On receipt all samples were tested using a nucleic acid amplification test—the ligase chain reaction (LCR) system (Abbott Laboratories) which was used according to the manufacturers instructions. To avoid any delay in transmission of results these were sent, on the day generated, to a secure fax machine at the relevant BPAS clinic—that is, the location at which the TOP would be performed. To maintain confidentiality, only the patient’s BPAS number, date of birth, and operation date were used as identifiers. Full printed reports followed by post.

Management of chlamydia positive women

All women with a chlamydia positive LCR result were managed in accordance with the study protocol which included treatment with a single 1 g dose of azithromycin. Positive results were generally given at the clinic, on the day of the TOP, by a clinic healthcare worker who also gave general information on chlamydia. All chlamydia positive patients were referred, with a letter, to their local genitourinary medicine (GUM) clinic for further management and partner notification.

The information pack provided for the patient included further information about chlamydia, two timetables of opening hours—one of which was for the partner, and introduction slips for the local GUM clinic. In addition, BPAS staff completed a checklist for each patient to ensure that the following were addressed: antibiotics given or referral for treatment, information given about chlamydia either verbal or written, importance of partner(s) being treated, discussion of no sex until partners are cured, referral letter to GUM department given.

Data collection, transfer, and analysis

A specially designed laboratory request form/questionnaire was used which was similar to that developed by some of us (JH, HM) for use in the NHS Chlamydia Pilot Screening Programme.⁶ The form consisted of a main portion (collecting core demographic and relevant clinical data) with a detachable key to response boxes for more detailed behavioural, service utilisation, and risk factor information. It was anticipated that all patient based data (clinical, demographic, and behavioural) required for the project would be obtained on the laboratory request form and that the detachable key would obscure sensitive information in transit. Data entry was carried out by appropriate staff at the testing laboratories. Data were initially stored in secure (password protected) laboratory database systems (TELEPATH )

Analysis was performed using stata version 6.0. Overall prevalence of C trachomatis infection (main outcome measure) was estimated from women testing LCR positive for infection as a proportion of all women providing a FCU sample. Age specific prevalence for C trachomatis was calculated. Testing site, demographic, and behavioural variations in chlamydia prevalence were also explored. The χ² test or Fisher’s exact test was used to explore associations between factors and whether a sample was provided and its result. Significant associations found in the single variable analysis with respect to the prevalence of C trachomatis infection were explored in a multivariable logistic regression model to determine the factors significantly and independently associated with being diagnosed with C trachomatis infection.

Ethics

Approval was obtained from the ethics committees of both BPAS and PHLS.

RESULTS

A total of 1293 women participated in the study and completed the questionnaire. Of these, 292 declined testing (84 because of a previous test) and 1001 (77%) provided a urine sample. Comparing women who accepted or declined the urine test, there were no significant differences with respect to age, ethnicity, or recorded partner change in the past year (see table 1). However, significant differences were found in their chlamydia testing and diagnosis history, symptomatology, and reproductive history. Women who accepted the urine test were less likely to have been previously tested for chlamydia (7.7% v 28.1%. p<0.001) and less likely to have had a previously positive chlamydia test (1.8% v 6.6%, p<0.001). They were more likely to be symptomatic (39.6% v 7.8% reporting at least one symptom, p<0.001) and were also likely to have had more children, miscarriages, and previous TOP.

Prevalence of C trachomatis

Three patients provided insufficient urine for LCR testing. Of the remaining 998, 75 (7.5%) were found positive for C trachomatis infection. At the four Midlands branches a total of 505 urines were tested and at the six north of England branches 493 urines were tested with positive rates of 7.5% overall in both regions. At the five branches which collected over 100 samples each, the positivity ranged from 5.8% to 8.2%. Chlamydia positivity was significantly associated with younger age (p<0.001), having previously tested positive (p=0.02), and having been previously treated (p=0.01) for genital chlamydia. Age specific prevalence ranged from 1.8% among those over 35 years to 11.5% among those aged between 20–24 years (table 2). Relatively high disease prevalence (10.8%) was also observed among those aged under 20 years. Seventy seven of the women tested reported a previous chlamydia test. Those who had a previous chlamydia test were equally likely to have chlamydia diagnosed compared with those who had never been tested previously (7.8% v 7.5%, p=0.9). However, women who had previously been diagnosed with chlamydia were significantly more likely to be chlamydia positive than those who had tested negative previously, as well as those who had never been previously tested (23.5% v 0% and 7.5% respectively p=0.02). A chlamydia prevalence of 20% (3/15) among women who had been previously treated for chlamydia raises questions about adequacy of treatment or reinfection.

We found no association between chlamydia positivity and ethnicity, nor with women’s willingness to pay for chlamydia testing. Neither was there any association between positivity and the presence of any reported clinical symptom(s), nor with the number of symptoms reported by the patient. A single question exploring sexual partner change and chlamydia positivity was included in the questionnaire; however, its completion rate was poor (65% provided a response). Chlamydia prevalence among those who reported having changed a partner in the previous year was 6.2% (7/113) compared with 7.6% (41/538) among women who had not done so. There were no significant associations between chlamydia positivity and numbers of children (p=0.8), number of previous miscarriages (p=0.6), and previous termination of pregnancy (p=0.2).

Factors associated with C trachomatis positivity

Those factors exhibiting evidence of an association (p≤0.2) with chlamydia positivity—age, ethnicity, previous positive test, previous treated infection, irregular bleeding, and the number of previous TOP—were included in a multivariable logistic regression analysis to find those factors independently associated with chlamydia positivity using a backward selection approach to obtain a final model. Four factors—age (p=0.002), ethnicity (p=0.01), previous positive test (p=0.06), and irregular bleeding (p=0.05)—remained in the model. All possible two way interactions between these factors were tested but none was significant.

The impact of charging for chlamydia screening

Women were asked whether they would have consented to a chlamydia test if they were to be charged £20 (table 3). Overall, 35% (range 8–51% across branches) of women who provided a urine sample indicated that they would still have requested the test.

The feasibility of C trachomatis testing in BPAS sites

A number of process measures were included to explore the efficiency and effectiveness of specimen transfer, testing, and reporting between BPAS sites and the public health laboratories. Measured indicators (table 3) included the proportion of test results reaching the clinic by the operation date (reporting delay) and the proportion of samples reaching the laboratory after the operation date (specimen transfer delay). Only 4.6% (range 0–70%) of samples arrived at the laboratory after the operation date and 86.2% (range 30–95.4%) of results arrived at clinics by the operation date.

DISCUSSION

This study examined the feasibility and outcomes of undertaking urinary screening for C trachomatis infection at BPAS branches and explored the possible impact of charging for this service. The high prevalence of chlamydial infection among attendees (11.2% in those below 25 years of age) confirms the vulnerability of this clinic population and is consistent with findings from other developed countries.^7–9 Examination of risk factors enables the identification of possible confounders of analysis for surveillance purposes (for example, age, previous diagnosis of chlamydia). Prospective monitoring of chlamydia positivity in this setting may be used to reflect the impact of chlamydia screening in primary care sites as advocated² by the chief medical officer (CMO), or other community based sexual health interventions. In a wider context, changes in chlamydia prevalence may parallel changes in general sexual health of a population allowing feedback to health authorities/primary care groups on the effects of local implementation of sexual health strategies. The level of sampling achieved suggests that representative and consistent data can be obtained.

Screening for chlamydial infection in women seeking termination of pregnancy has been recommended² and it will be important to develop a framework through which the prevalence of C trachomatis infection among service users may be monitored. Our study highlighted the feasibility of coordinating chlamydia screening within the current TOP procedures. We are encouraged that very few samples experienced significant transport delay to the laboratory and that the majority of clinics received the results in time for the procedure. Clearly, any proposed screening programme within BPAS sites would need to be funded and one strategy would be to ask women themselves to pay for the chlamydia testing. Our study suggests that many women would be unwilling or unable to do so. Just over a third of women would be willing to have the chlamydia test if they were requested to pay £20. Although there were no recorded significant differences between those who were willing to pay and those who were not, failure to screen 65% of women would be unsatisfactory for both public health and prevalence monitoring purposes. We would strongly encourage that chlamydia testing is offered free of charge to patients utilising the services of BPAS. The public health benefits equate to those from the wider screening of the population, as recommended by the CMO, and should be funded similarly.

A key consideration for successful implementation of a TOP clinic based screening programme is its acceptability to staff. Although not formally evaluated in this study, we obtained and documented the informal feedback of clinical colleagues who were involved in the implementation and delivery of this study. BPAS staff preferred the model used—namely, urine testing at the branch with results available at the clinic, to other possible alternatives that included near patient testing at the clinic (some clinics had no on-site laboratory capability and there were time constrains in having the result available), or sending swabs to the microbiology laboratory at the time of the operation (no opportunity to discuss results and implications with the patient).

There are a few limitations to our study. The lack of association between recent sexual partner change and chlamydia positivity, in our study, in part reflects the poor completion of this item by clinic staff. This resulted from some misunderstanding and embarrassment experienced by patients and clinic staff. Secondly, although every effort was made to ensure that results were received at clinics before the TOP date, this was not possible in some cases. The poorer performance figures relating to tests undertaken at branches B and H were due to misunderstandings in transport arrangements and could easily be addressed. The low (55%) achievement for results reaching the clinic by the day of operation for patients tested at branch E reflects the higher proportion of patients who can be offered an operation date closer to their consultation visit at that particular BPAS site. Finally, all patients diagnosed with chlamydia in our study were referred to a GUM clinic in case of other concurrent infections and for partner notification (a key secondary prevention activity aimed at reducing onward transmission and potential complications of undiagnosed disease) but we did not collect information on these outcomes.

This evaluation focused on the feasibility of undertaking routine chlamydia screening among BPAS attendees. We conclude that a BPAS/PHLS collaboration would make it feasible to monitor chlamydia prevalence in this nationwide service while offering a screening intervention that is likely to bring significant benefit to patients, their partners and, indirectly, to the wider population. More detailed examination of the potential psychosocial benefits and costs of chlamydia screening in this population, and strategies for ensuring effective partner notification and treatment, should now be explored.