METHODS

Population studied

This study took place in an outer London GUM clinic, which sees approximately 16 000 patients a year, 2% of all UK gonorrhoea, and in which 27% of all attenders are of white UK ethnicity. All new patients attending the GUM service during the two study periods were asked to complete a self completion questionnaire, and computerised clinic data for all patients were entered into the analysis. Patients who had ever had an HIV diagnosis recorded at the clinic accessed a different service within the clinic, and were not included.

All categories of staff working at the clinic during phase 2 received a questionnaire asking their views on the two access systems.

Access policies before and after the change

Until the end of March 2001, the clinic operated a wholly walk-in service, in which all patients arriving before the end of a clinic session could wait to be seen. No triage or selection operated, and waiting times varied up to approximately 3 hours.

After 1 April 2001, all clinic sessions consisted of approximately 35% prebooked appointments, and 65% slots bookable only on the day of attendance. This ratio was chosen to ensure that the number of same day slots would exceed the number previously used by new patients or patients with a new problem, thus allowing for at least the same number of people with an acute problem to have same day access, through a numbers limited walk-in service. Same day slots could be booked in person only, in order to ensure full usage, and maximise use by patients highly motivated by acute problems. Nurse triage was offered to attenders who could not be allocated an appointment on the day and wished to be seen immediately, and the triage policy required that young people (under 17), those with symptoms suggesting an acute STI, sexual assault cases, and those with a contact slip should be seen during the current session. Interviews with reception staff revealed that when the day’s slots were full, they advised patients with “tips” on how to be sure of getting one of the next day’s slots, and that this may have reduced demand for triage.

Phase 1 of the study was the calendar month immediately preceding the change, and phase 2 was an 8 week period beginning 10 weeks after the change. Neither period included a prolonged holiday weekend closure. No other relevant significant change took place in the clinic during this entire period.

Study design and data collection

Figure 1 gives an overview of the study design. All patients attending the clinic during phase 1 and phase 2 were asked to complete a two page questionnaire (see Appendices 1 and 2 on STI website) which was given to them by reception staff and collected in a closed box. This explored health seeking behaviour, use of other services, sources of information about the clinic, preferences and use of appointment types, and likely actions had they not been able to access services on that day. The questionnaire was labelled with the date and patient number. An information sheet explained that responses would be linked to clinic data. Brent medical ethics committee approved the study. Numerical data were double entered onto a database, and anomalies resolved by checking against the original.

Clinical computerised data, including age, ethnic group, present and past STIs were downloaded in disaggregated form, into a single database. This was used to construct variables relating to past and present attendances, and a variable indicating diagnosis of an acute STI during the relevant study period. The latter was used to select case notes for individual review. Routine computerised quarterly surveillance data for the clinic were also collected, measuring the number of episodes (attendances for a new problem, and overall attendances).

Case note review was undertaken by a single reviewer (JC) in order to ascertain symptom status, the success or failure of partner notification as judged by clinic staff, and the achievement or non-achievement of follow up. Partner notification was assumed to be required only for bacterial STIs including pelvic inflammatory disease and epididymo-orchitis. Successful partner notification was defined as the attendance and treatment of at least one partner to the documented satisfaction of the health adviser or clinician seen on follow up attendance.

The clinical, questionnaire, and case note review data were merged after their completion, in order to allow linkage of questionnaire responses to clinical history. Data were also collected on attitudes to and use of primary care, symptom duration, and views on the need for treatment. These results are not presented here.

Statistical analysis

The χ² test, and, where appropriate, the t test for difference between means, were used to make comparisons between the following groups: phase 1 v phase 2 attenders and clinical outcomes; new patients using slots available on the day v new patients using prebooked appointments; preference for prebooked appointment or walk-in by phase and by patient characteristics; characteristics of responders v non-responders.

RESULTS

The overall clinic population did not change between phases 1 and 2. Table 1 shows that the proportion of patients with any STI and with specific STIs, mean age, and proportion of new patients did not change between phases. The ethnic profile of the population was also unchanged, as was the proportion of patients who were attending with a new problem to all attendances including follow ups (53.1% in phase 1, 53.7% in phase 2). However, the proportion of men increased slightly in phase 2, just reaching statistical significance. Total attendances in phase 2, compared to the same quarter of the preceding year, increased from 3610 to 3966, while the mean number of attendances by patients with a new problem was 38.0 per day in phase 1 and 37.9 per day in phase 2. The median time from first contacting the clinic to being seen was 1 day in both phases, for walk-in patients and patients using a slot available only on the day.

There was some evidence of improved clinic outcomes in phase 2, with a higher uptake of HIV testing among new patients (p=0.04), and a trend towards higher partner notification rates for acute bacterial STIs (table 1).

The questionnaire was completed by 700 (69.4%) of 1084 new patients, with no difference in the response rate between phases. Respondents were slightly older, and less likely to be male or of Afro-Caribbean ethnicity than non-responders (table 2). Preferences are summarised in table 3. The majority of patients preferred a walk-in service. This preference weakened but remained in phase 2, and was most pronounced in patients attending because of symptoms, males, and patients under 25.

The use of appointment slots by new patients is analysed in table 4. Data on appointment type are missing for some dates in phase 2, owing to accidental destruction of clinic records. Only new patients are analysed, since new patients’ access to clinics is of particular concern, and the questionnaire was administered only to new patients. Patients under 25 years and non-white patients were less likely to use prebooked appointments. However, this was not explained by symptom status, since the proportion of patients attending because of symptoms was similar between the two groups. The proportion of patients diagnosed with an STI did not differ between appointment and non-appointment users.

All staff working in the clinic just after phase 2 completed the staff questionnaire, and 11/12 (91.6%) staff who had experienced both phases of the study preferred the new access policy. The most commonly cited reason was a more even workload, followed by a view that patient care was better in the new system.

DISCUSSION

The natural experiment observed in this study took place against the background of increasing numbers of STIs diagnosed in the United Kingdom in recent years. This busy outer London clinic, which previously saw all patients arriving during clinic opening hours, was finding it difficult to cope with increasing numbers of walk-in patients. There were problems of variable patient flow across the day, long waiting times, aggression among patients, and high levels of stress for staff. As a result, it was decided to switch to the limited appointment system we have described.

The design of this study assumes that walk-in clinics, the traditional access system for GUM clinics, provide optimal access to patients, and this is the benchmark against which the phase 2 results are compared. The majority of clinics in the United Kingdom are no longer primarily walk-in, and there is currently a lack of evidence for the effect of recent changes in patient access policies. Nevertheless, in the absence of other evidence an appointments system should demonstrate equivalence of access with a walk-in service. Our results demonstrate that overall equivalence of access to a walk-in service can be preserved while using 35% of prebooked appointments, without loss of at-risk patients or minority ethnic groups.

HIV testing rates and partner notification rates improved in phase 2, probably as a result of the smoother workload allowing better allocation of time on the basis of clinical need. A staff questionnaire suggested that all groups of clinical staff felt they were providing better care to patients in phase 2.

However, young patients, Afro-Caribbeans, and Asians relatively underutilised appointments, while expressing strong preferences for a walk-in system. This a matter of concern, and suggests that while our phase 2 mix of prebooked and “on the day” slots may preserve access, there is a risk that a further move toward appointments would be expected to disadvantage these groups, who may be at increased risk of STIs.^12,13 Preference for a walk-in service decreased from over three quarters to half of new patients in phase 2, when appointments became available for the first time. It is possible that not all patients in phase 2 were aware that prebooked appointments were available.

Our study was located in an ethnically mixed, socially deprived district where access to STI services appears to be good (as measured by the short median time from contacting the clinic to being seen). Our results should not be assumed to generalise to all mixed walk-in/appointment systems, particularly where there is a smaller proportion of slots available on the day.

The survey was limited to patients new to the clinic. As a consequence, we lack behavioural and attitude data on patients who had previously attended, and may be a high risk group whose health seeking behaviour should be studied in future work. We also do not have accurate data on the number of patients turned away, and whether they were seen the next day, nor do we have data on the number of patients triaged. This would be useful information, and will be collected in any follow up of this study.

Two other clinics are currently using this methodology to study the effects of changing from walk-in to different forms of appointment system, and this will provide useful empirical data for clinicians planning to change services. If an evidence base can be built up on the effects of running different kinds of appointment in the specific context of GUM services, it should be possible in future to make evidence based decisions on access structures, with a view to maximising access for those most at risk. Given the small numbers of walk-in clinics remaining, it is important that opportunities to evaluate the move towards appointment based services are taken quickly. Other changes in access policy should also be evaluated, and it is hoped that the methodology we have described will provide a useful, comparable tool for assessing a variety of changes in the access structures of sexual health services.