Rationale and development of a survey tool for describing and auditing the composition of, and flows between, specialist and community clinical services for sexually transmitted infections

Abstract

Background

National health strategies have called for an expansion of the role of primary care in England to increase access to sexual health services. However, there is little guidance for service planners and commissioners as to the public health impact of different combinations of specialist genitourinary medicine (GUM) clinics and primary care based services for local populations. Service planning for infectious diseases like sexually transmitted infections (STI) is further complicated because the goal of early detection and treatment is not only to improve the health of the individual, but to benefit the wider population and reduce future treatment costs by preventing onward transmission. Therefore, we are developing a survey tool that will enable service planners to better understand the needs of their local STI care-seeking population and which will help inform evidence-based decision-making about current and future service configurations. Here we describe the rationale and development of this survey tool.

Methods/Design

A pen-and-paper questionnaire asking about sociodemographics, reasons for attendance, care pathways, and recent sexual risk behaviours, is being developed for patients to complete in waiting rooms of diverse clinical services, including GUM clinics and primary-care based services in sociodemographically- and geographically-contrasting populations in England. The questionnaire was cognitively tested before being piloted. In the pilot, 67% of patients participated, of whom 84% consented to our linking their questionnaire to data on STI testing and diagnosis and partner notification outcomes from their clinical records.

Discussion

The pilot study suggests that both the questionnaire and its linkage to routinely-collected clinical data are likely to be acceptable to patients. By supplementing existing surveillance, data gathered by the survey tool will inform service planners' and providers' understanding of the needs and care-pathways of their patients, facilitating improved services and greater public health benefit.

Background

Testing and treatment for sexually transmitted infections (STIs) in England has historically largely taken place in genitourinary medicine (GUM) clinics, although an increasing amount is occurring in primary care.[1] This reflects the recommendations of the National Strategy for Sexual Health and HIV[2], and more recently, the MedFASH/BASHH Standards for the Management of Sexually Transmitted Infections (STIs)[3] call to expand the role of primary care further to increase access to sexual health services. As a result, there has been growth in more specialised but highly variable[4, 5] models of Local Enhanced Services for Sexual Health (LESSH) in primary care.[6] However, the published recommendations fail to provide commissioners and planners of sexual health services guidance as to how to decide on the relative capacity and characteristics of these clinical services to meet the sexual health needs of local populations, making local commissioning difficult.[4, 5] Service planning for infectious diseases like STIs is further complicated because each case can produce further cases,[7] so the goal of early detection and treatment is not only to improve the health of the individual but that of the wider population by preventing onward transmission, which also reduces future treatment costs.[8] Cost-efficient services therefore need to provide rapid and appropriate care, tailored to the needs of their local populations.

The MSTIC study, an abbreviation of 'Maximising STI Control' (full study title: 'Public health outcomes of GUM and primary care-based STI services: How to maximise STI control for a population'), is a UK Medical Research Council funded study (grant number G0601685). The MSTIC study aims to develop an evidence-based, web-based tool to assist those planning sexual health services for local populations in determining the relative public health impact of different combinations of health services. The web-tool will incorporate the results of a discrete event simulation mathematical model of the key factors influencing the transmission of common STIs so that the effect of different combinations of clinical services on averting transmission can be assessed. The model will use publicly-available data about local populations, for example: local census data and routinely-collected surveillance such as the GUM Clinic Activity Dataset (GUMCAD)[9] and GUM Access Monthly Monitoring (GUMAMM) data.[10] However, while GUMCAD and GUMAMM can provide basic sociodemographic data as well as data on STI testing and positivity for patients attending GUM clinics and increasingly, primary-care based LESSH services, the range of relevant information collected by these surveillance systems is limited.

We are therefore developing a survey tool in the form of a patient questionnaire that can be linked to an extract of patients' clinical records for clinical services to use to provide much more insight into their local patient populations, including questions on their care pathways and transmission-risk behaviours. Data collected by the survey tool can then be used to inform evidence-based decisions about service configuration either via the MSTIC web-tool, or independently, in the context of audits and/or service evaluation, enabling data to be gathered that are comparable over time and between services. This paper now describes the development of the survey tool in GUM clinics and primary-care based services.

Ethics

The survey tool is being developed in a research context and as such has required Research Ethics approval, which was obtained from the London Research Ethics Committee (reference: 09/H0718/1). However, the tool is intended for routine use in audit and service development, which do not currently require ethical review in the UK[11, 12].

Methods/Design

Design

We are using a quantitative, cross-sectional survey design to collect data via a pen-and-paper questionnaire from patients attending clinical services offering care for STIs. Data from patients' questionnaires are then linked, with patients' consent, to an extract of their routinely-collected clinical data by means of patients' clinical identifiers.

Settings

Three sociodemographically- and geographically-contrasting areas in England, broadly representing urban, suburban and rural populations, have been purposively selected to implement the survey tool.

Study population

The study population is defined as patients attending four principal GUM clinics and, in the rural area, a further four satellite clinics operating from the main GUM clinic on a weekly or fortnightly basis. Patients attending three general practices operating as LESSHs by each holding a weekly sexual health clinic session in the rural area, are also being surveyed.

Acknowledging that young people are a particular 'risk group' for poor sexual health[13, 14] we have decided to include teenagers so that we can capture the needs, access patterns and services received by this important minority of service users. However, we are excluding children aged 13 and under as they form a very small minority of sexual health service users, for which ethical issues around informed consent to participate in research are likely to outweigh the advantages gained through including the likely small number of patients of this age.

Study materials will only be available in English as the participating services report a high level of literacy in English in their non-UK born patients, and also due to feasibility constraints.

Implementation

Reception staff at the services are to offer questionnaires to patients upon arrival at reception and to mark each questionnaire with the date of attendance and the patient's clinical identifier. In the GUM services, questionnaires are for all patients, whereas in the LESSH services run in general practices, questionnaires are administered during sexual health clinic sessions, meaning that sexual health patients can be identified without asking their reason for attendance.

All patients are asked to tick a box on the first page of the questionnaire to indicate consent to our linking their questionnaire data to an extract of their clinical records. Upon the study co-ordinator (CA) receiving completed questionnaires, the services will be provided with a list of the clinical identifiers and dates of attendance of consenting patients. Depending on the service's preference, services will either provide an electronic extract of the relevant data from their clinical database, or will be provided with paper forms or spreadsheets in which to record extracts of the clinical data of consenting patients.

Development of data collection instruments

The data collected by the questionnaire include basic sociodemographics, reasons for attendance, duration of care-seeking and any experience of other services, recent sexual history (including sex since recognising a need to seek care), and presence and duration of symptoms. The questionnaire is based on that used by a previous study,[15] enhanced to include questions about recent sexual partnerships, adapted from the forthcoming third British National Survey of Sexual Attitudes and Lifestyles (Natsal-3),[16] a national probability survey of sexual behaviour. In particular, we are seeking to distinguish partnerships that have ended from those that are ongoing when the patient attends the service, as this has implications for the likelihood of successful partner notification.[17] Data on STI tests performed, STI diagnoses made, and partner notification outcomes will be obtained from the clinical records of consenting patients. A full list of data items, including the rationale for collecting each item, and its source by service type, is presented in Table 1.

Table 1 Data items collected by the survey tool by source and service type and rationale for collection.

Though collecting comparable data, the questionnaire and clinical data extract differ slightly between GUM and LESSH services, reflecting differences in the nature of the services and in the patient data routinely collected for surveillance purposes. As such, LESSH patients will be asked a different question about their past use of the service to distinguish between attending the practice for sexual health reasons versus other reasons. If registered with a GP, these patients will be asked which practice, for the purpose of determining whether they are registered with the LESSH practice they are attending, another LESSH service, or a practice which does not provide this service. Recognising that staff in general practice are not primarily sexual health specialists, the LESSH data extract includes treatment administered, while we assume that clinical guidelines (e.g. [18, 19]) are followed in GUM clinics. GUM clinical data extracts include patients' PCT of residence as people often attend clinics outside their PCT of residence, especially within London[20]. There are some further minor differences in questionnaire and clinical data extract wording between versions - see Additional files 1, 2, 3, 4, 5 and 6 for the questionnaires and the paper proforma for clinical data extracts for use by each service type.

We carried out a two-stage pilot to test the questionnaire. Thirteen cognitive interviews with patients at one of the participating GUM clinics were undertaken to assess whether the survey tool was acceptable to patients, whether questions were understood and answered as the research team intended, and whether response options were adequate. Data on the sociodemographic characteristics of these patients are provided in Table 2 and were considered by clinic staff to be broadly representative of their clinic population. The second stage involved piloting the questionnaire over a two-day period at the same clinic, which resulted in 56 out of 81 patients completing the questionnaire, i.e. a 67% response rate, of whom 84% consented to linkage. Questionnaires in this pilot were completed with low item non-response, very few inconsistent answers, and instances where the respondent had not followed the routing were minimal.

Table 2 Sociodemographic characteristics of patients participating in cognitive interviews

Patient confidentiality

Precautions taken to protect patient confidentiality include: not asking for patients' names, addresses or signature at any point - instead asking patients to tick a box to indicate consent; providing locked boxes in reception areas so that other patients do not have access to completed questionnaires; and providing envelopes in which patients seal their questionnaires, so that patients can conceal their questionnaire if they move around the clinic and so that clinic staff outside of the research team do not see completed questionnaires. The questionnaire data and clinical data extracts will be pseudonymous as both datasets contained patient clinical identifiers, necessary for linkage. Nonetheless the databases will be securely stored in password-protected files and after linking the datasets, the clinical identifiers will be deleted, rendering the data anonymous.

Data collection periods

Data collection is likely to take from three to ten weeks per service, according to the likely numbers of patients attending the service and anticipated response rates, based on experience from an earlier study.[15] Sample size calculations are not considered necessary, reflecting the aim of gathering data rapidly to inform service planning, and that use of the survey tool should, as an audit of the patient populations attending services, provide indicative data, such that the attainment of large numbers will not be necessary.

Data entry and analysis

Questionnaire data and clinical data recorded on paper forms will be double-entered into a Microsoft Access database by study administrators (estimated to take approximately five minutes per questionnaire, two minutes per data extract). These data will then be transferred into the statistical package STATA version 10[21] for statistical analysis. However, users of the survey tool without access to a specialist statistical package will be able to use the query functions in Microsoft Access to generate descriptive statistics. Any sub-group analyses with denominators of less than 25 will need to be treated with caution and 95% confidence intervals will be reported with all estimates to reflect the effect of the size of the achieved sample.

Discussion

To our knowledge, this will be the first study to collect comparable, detailed data from specialist GUM clinics and primary-care based LESSH services. To date, little has been known about the characteristics and behaviours of patients attending new models of sexual health service provision, such as LESSH. Our survey tool will provide a feasible and rapid means of collecting data from diverse clinical services, including these highly variable[4, 5] models of sexual health service provision.

We anticipate satisfactory response rates under the conditions in which clinical services are likely to use the survey tool for planning purposes: researchers will not be present to administer the questionnaire, there will be no incentives or tokens of appreciation for patients or reception staff, and the questionnaires cover some highly-sensitive topics. The response rate for the pilot study was similar to the 65.4% response rate achieved by the last British probability survey of sexual behaviour (the second National Survey of Sexual Attitudes and Lifestyles),[22, 23] a researcher-intensive study. However, there may be considerable variation in response rates between services, which was observed in the earlier study and attributed to enthusiasm for the research varying between reception staff teams, in some services reflecting short-staffing.[15]

We will only be able to administer the LESSH survey tool in one of the three study areas for reasons of feasibility and appropriateness: one area currently does not have a LESSH and in the other, patients seeking care for suspected STI are currently seen within the general surgery lists and we feel that it will be inappropriate for receptionists to identify these patients by asking them their reasons for attendance. However, should such services wish to implement the survey tool then the questionnaires could be distributed during patients' consultations, requiring reliance upon the co-operation and memory of GPs and practice nurses instead of reception staff.

Ensuring enthusiasm from those staff charged with distributing the questionnaire is expected to be important in terms of achieving both a high response rate from patients and a high rate of accurate completion of patients' clinical identifiers and dates of attendance on the front of the questionnaires. While we anticipate a high proportion of respondents will consent to our linking their questionnaire data to their clinical data, in a previous study[15] this was not possible for a handful of cases as clinical identifiers and/or dates were illegible, missing, or the clinical identifier did not correspond to an attendance on the date given. Aside from time, the enthusiasm and organisational skills of key individuals, such as reception supervisors, is likely to be vital to the successful administration of the survey tool.

Clinical services that have used electronic patient records for some time should be able to provide the research team with their clinical data extracts quickly. However, for services that still rely on paper notes this is likely to be a time-intensive process. Obtaining the clinical data extract will become quicker and easier as all clinical services move towards electronic patient records systems, and this will further facilitate the use of the survey tool.

In conclusion, development work for our survey tool suggests that it will be a rapid and easy method for sexual health services to collect detailed, individual-level data on their patients, supplementing existing surveillance data to better understand their patient populations and their care pathways so that services can be made more clinically- and cost-effective. Our paper's additional files will allow service providers and planners to implement the survey tool themselves, which will be feasible with enthusiastic staff, without the need for researchers to be present or additional resources.