Audit of management of acute infective conjunctivitis in primary care
Locum GP
E-mail: gayle.knights{at}btopenworld.com
 |
Abstract |
|---|
 TOPNotes Abstract Criteria chosenStandard setPreparation and planningData collection 1Changes to be evaluatedData collection 2ConclusionsReferences |
|---|
A recent article in the British Medical Journal (Everitt et al., 2006), looking at various options for the management of uncomplicated acute infective conjunctivitis (AIC), suggested that a delayed prescription of anti-infective eye preparations was the most appropriate choice. Up to this point, I routinely prescribed chloramphenicol or fusidic acid for all patients presenting with AIC and the partners at my practice agreed their management approach was the same as mine. A literature search of Medline, Cochrane Library and BMJ.com, using search strings ‘infective conjunctivitis’, ‘delayed prescribing’ and ‘conjunctivitis-delayed prescribing’ revealed further evidence for withholding treatment.
Acute infective conjunctivitis (AIC) accounts for up to 2% of primary care consultations (Everitt and Little, 2002; Rietveld et al., 2004) and in England 3.4 million community prescriptions are issued each year for anti-infective eye preparations (Rietveld et al., 2004). Only 50% of cases are bacterial in origin (Everitt and Little, 2002; Rietveld et al., 2004) and clinically it is very difficult to distinguish these from viral causes (Rietveld et al., 2004).
Further evidence suggests that 64% of cases of AIC resolve spontaneously within 2–5 days without treatment (Everitt and Little, 2002). More importantly, sight-threatening complications are rare (Sheikh and Hurwitz, 2006).
The first major trial looking at management of AIC was carried out in Oxfordshire (Rose et al., 2005) and compared treatment with chloramphenicol eye drops versus treatment with a placebo. Their results showed that cure rate, mean time to cure, relapse rate and adverse event rate was the same in both treatment groups. The strengths of this trial were the numbers involved (326 children in total) and its randomized double-blind nature. Its main weakness is the use of placebo, which could be argued to be providing a lubricating effect.
The second major trial was the Hampshire, Wiltshire and Dorset trial (Everitt et al., 2006). This compared treatment with immediate antibiotics versus delayed prescribing (prescription collected from surgery after 3 days, if required) versus no antibiotic. The results of this trial were interesting for two reasons. Firstly, it showed that using delayed prescribing significantly reduced uptake of antibiotics (99% in the immediate antibiotic arm vs. 53% in the delayed prescribing arm) with little prolongation of moderate symptoms (mean 3.3 days in immediate antibiotic group vs. 3.9 days in delayed antibiotic group).
Secondly, the authors looked at patients’ health beliefs after treatment demonstrating decreased belief of antibiotic effectiveness and decreased intention to reattend with a similar problem, for patients in both the delayed prescribing group and the no antibiotic group, when compared with the immediate antibiotic group. The strength of this trial is the size (307 patients total) but its weakness is its inability to be blinded: both the patient and the doctor knew which treatment they received, introducing possibility of selection bias. Some critics have also said that using a delayed prescribing strategy increased antibiotic uptake by 23%, because 53% of patients in the delayed prescribing arm collected their prescription, versus 30% in the no antibiotic arm receiving later antibiotics.
Our GPs were concerned that delayed prescriptions do not decrease antibiotic uptake. In fact, evidence shows that less than 50% of delayed prescriptions are used with largest reductions in uptake occurring if the patient must return to the surgery to collect (Little et al., 1997, 2001; Dowell et al., 2001; Arroll et al., 2002, 2003; Cates 2003).
There are benefits to be had from avoiding prescribing in minor, self-limiting illnesses:
- Reduction in costs—£5.8 million was spent on anti-infective eye preparations in the UK in 2005 (Department of Health 2005)
- Reducing medicalization and reconsultation for minor, self-limiting illnesses (Dowell et al., 2001; Everitt et al., 2006)
- Antibiotic resistance is on the rise and we are encouraged to prescribe responsibly—I spoke with our local microbiologist and discovered that we are starting to see some resistance in our area to fusidic acid
We do still need to consider our patient population's expectations which, based on their previous experiences, may present some challenges. Two qualitative studies (Everitt et al., 2003; Rose et al., 2006) looked at patient ideas.
- Most patients are confident to self-diagnose AIC and see it as a minor condition
- Most were unaware that it was self-limiting and assumed that antibiotics should be started as soon as possible—they had not considered the consequences of leaving it untreated
- About 20% of educational institutions would not allow children back if they were infected or were currently using antibiotic eye drops—a further 31% would allow children back once started on eye drops but required the parent to administer the medication at the school
- Patients welcomed the news that conjunctivitis is self-limiting and preferred not to receive antibiotics if they could avoid them
- Most preferred a delayed prescribing approach to no antibiotics at all as it allowed a ‘safety net’ with minimal inconvenience to doctor and patient
The practice discussed the above evidence at a multidisciplinary meeting and agreed that it would be appropriate and safe to change our prescribing in AIC. We discussed the possibility of not prescribing at all, but felt this had the potential to increase workload, as 30% would reconsult for antibiotics (Everitt et al., 2006). We then agreed to audit our management and see if our change was successful. Our agreed audit question is ‘Are patients with AIC being managed with a delayed prescription’?
|
Criteria chosen |
|---|
|
TOPNotesAbstractCriteria chosenStandard setPreparation and planningData collection 1Changes to be evaluatedData collection 2ConclusionsReferences |
|---|
Based on the previous research, we agreed as a practice to the following audit criterion:
- Qualifying patients presenting with AIC should be managed with a delayed prescription
As the two main trials in this area (Rose et al., 2005; Everitt et al., 2006) had specified exclusion conditions, we felt it appropriate to adopt those same exclusions, as below:
- Age less than 6 months
- Symptoms present greater than 5 days
- Systemically unwell
- Evidence of severe infection, for example peri-orbital cellulitis
- Immunocompromised
- Allergy to any anti-infective eye preparations
- Taking any antibiotic in the previous 48 hours
- Chronic infective eye disease
- Eye surgery in the previous month
|
Standard set |
|---|
|
TOPNotesAbstractCriteria chosenStandard setPreparation and planningData collection 1Changes to be evaluatedData collection 2ConclusionsReferences |
|---|
It was difficult to set an appropriate standard for this audit due to a lack of comparative work. I spoke with our PCT pharmacist, local audit liaison and my colleagues at the Day Release Course but was unable to find any appropriate audits. A practice meeting, with the doctors, was held to agree a realistic target.
The general consensus was that we would currently be performing poorly due to previous habit, patient expectation and inexperience with delayed prescribing. We all recognized that it takes time to ‘teach an old dog new tricks’ and would, therefore, not wish to set a standard of 100%.
However, AIC is easily recognized and our exclusion criteria are very specific, making identifying appropriate cases relatively easy. For this reason, we felt we should still aspire to a reasonably high standard. Suggestions from the team varied from 60% to 95% and we took an approximate average, setting an agreed audit standard of 80%.
|
Preparation and planning |
|---|
|
TOPNotesAbstractCriteria chosenStandard setPreparation and planningData collection 1Changes to be evaluatedData collection 2ConclusionsReferences |
|---|
Our practice is in a rural setting, covering 200 square miles. We operate from a purpose-built health centre and have a branch surgery located 8 miles away. There are 3.5 whole-time equivalent partners at the practice, one doctor retainer, one nurse practitioner and a regular locum. Our list size is just over 6000 patients and we provide dispensing services to approximately half of these. Our patients tend to be affluent and well motivated. We use EMIS for our computer records.
As a first step to planning the audit, I did a computer survey to estimate how many patients we see with AIC (EMIS code F4C0). From this survey, I was able to see that we diagnose far more cases of AIC in the winter months than during the rest of the year.
In order to achieve a large enough data set for a full audit cycle, within the short space of time allotted, it seemed sensible to make the first data collection retrospective from the previous winter with a second data collection occurring prospectively during winter 2006–07.
In order to collect and analyse the audit data, a search was performed on the EMIS computer system, using the following parameters:
- All patients in database (shared feature)
- Diagnostic code F4C0—acute conjunctivitis unspecified and all its children, with the exception of acute atopic conjunctivitis
- Diagnostic code F4C0 to apply between given date ranges. For the first data collection, this would be 15/11/05–15/2/06 and for the second data collection 15/11/06–15/2/07.
As we have no code available for delayed prescriptions, we discussed our current recording methods within the practice. Free-text entries were variable, meaning each patient record would have to be reviewed individually, requiring a significant amount of time. It would also require to be performed by somebody with clinical training as a certain amount of interpretation would be necessary. Given that this is unlikely to become a regular audit and did not require a designated lead in the future, I was happy to perform the data collection myself.
In order to be as thorough as possible, I checked the following information in every computer record:
- Allergies recorded (Function ‘B’)—specifically to anti-infective eye drops
- Past problems (Function ‘P’)—suggestive of recent eye surgery or that patient is immunocompromised
- Medication (Function ‘M’)—recent antibiotics and medication affecting immunity
- Record of conjunctivitis episode was then reviewed for further details
This was time consuming but necessary as most relevant information was entered in free text.
Analysis of the data was kept as simple as possible, using the following five questions:
- Based on the consultation documentation, does this patient fit the criteria for inclusion in the audit—Yes or No?
- If not, what is the reason?
- If they do fit the criteria, did this patient receive a delayed prescription—Yes or No?
- What percentage of patients included in the audit received a delayed prescription?
- Was this percentage greater than, or equal to, the audit standard?
|
Data collection 1 |
|---|
|
TOPNotesAbstractCriteria chosenStandard setPreparation and planningData collection 1Changes to be evaluatedData collection 2ConclusionsReferences |
|---|
The EMIS search produced a list of 47 patients with diagnostic code F4C0 between 15/11/05 and 15/2/06 (search was performed on 7/11/06).
Characteristics
- Total patients identified = 47
- Patients excluded from audit (Table 1) = 12
- Remaining patients (included in audit) = 35
- Male = 12; female = 23
- Age range = 7 months–88 years
- Mean age = 32 years
Table 2 shows that <9% patients were treated with a delayed prescription or no prescription during the first data collection. This is much less than the audit standard.
|
|
|
Changes to be evaluated |
|---|
|
TOPNotesAbstractCriteria chosenStandard setPreparation and planningData collection 1Changes to be evaluatedData collection 2ConclusionsReferences |
|---|
The first data collection demonstrates that we are performing far below our audit standard. The reasons for this low performance are obvious and would continue to present obstacles unless addressed properly.
The data collection is retrospective and covers a time period that predates the British Medical Journal article (Everitt et al., 2006). During this time, the prescribing practitioners were using the best clinical knowledge available to them and had not considered a delayed prescribing strategy in these patients.
Prior to this audit, the practice did not have a system in place to enable delayed prescribing, which was used on an ad hoc basis by individual practitioners and usually involved the patient ringing the GP on a later date to request a script.
As a practice that does not routinely utilize delayed prescriptions, our patients are not familiar with the concept and are poorly equipped to understand the reasoning behind it. In addition, our local schools and nurseries adopt the policy that children should be excluded until they have been started on anti-infective eye preparations. These two factors together put a lot of pressure on the GP to prescribe.
The easiest of these problems to address was education of the prescribing practitioners. This was achieved by holding an educational practice meeting and presenting the evidence.
With this new information, the practitioners were asked to embark on some patient education during consultations for AIC. This education would include the facts about treatment of AIC, explanation of delayed prescribing and encouraging patients to utilize this new service.
In order to aid compliance, I also wrote a patient information leaflet (PIL; The information leaflet given to the patients or carers is available as supplementary data online) for patients receiving a delayed prescription. This leaflet was based on the PIL available on Prodigy (2006) but was edited to include data from my research, information about delayed prescribing and details of how our patients could obtain their prescription if needed. In addition, I wrote a generic letter (The letter for patients used in this audit is available as supplementary data online) for parents to give to their child's school/nursery explaining our new protocol for the management of AIC and the reasoning behind it.
The biggest challenge was to initiate a procedure for delayed prescribing within the practice. As a group, we agreed that patients should be asked to return to the practice for their script after 3 days, as in the British Medical Journal article (Everitt et al., 2006), if their symptoms were not improving. This would enable us to track prescription usage for ourselves and see if delayed prescribing was successful in our particular community.
It was important to us all that our system could be used for both dispensing and non-dispensing patients, without significantly increasing workload within the dispensary or disrupting stock level records.
The practice manager, dispensary manager and I held a meeting to discuss possible options, suggested previously by the partners.
- Diagnosing GP to record ‘delayed prescription authorized’ on notes and doctor on-call to issue on request. This seemed inappropriate for the busy on-call doctor.
- Diagnosing GP to record delayed prescription authorized on notes and patient to ring back the same GP. This would prove difficult in our practice as four GPs work part-time, so what would happen on their days off?
- Diagnosing GP to print and sign a prescription in all the cases and record ‘delayed prescription’ in case notes. Prescription to be held in dispensary for patient to collect if needed. This would cause problems for our dispensing patients.
Our dispensary manager selected option 3, feeling it would be easiest for both doctors and patients. Non-dispensing patients could be given their script on request, without any further inconvenience. For dispensing patients attending for their script, the dispenser could cancel the last electronic issue of the prescription and reissue for dispensing purposes. This would allow stock levels to remain accurate.
|
Data collection 2 |
|---|
|
TOPNotesAbstractCriteria chosenStandard setPreparation and planningData collection 1Changes to be evaluatedData collection 2ConclusionsReferences |
|---|
The second EMIS search produced a list of 32 patients with diagnostic code F4C0 between 15/11/06 and 15/2/07 (search was run on 6/3/07).
Characteristics
- Total patients identified = 32
- Patients excluded from audit (Table 3) = 13
- Remaining patients (included in audit) = 19
- Male = 11; Female = 8
- Age range = 6 months–81 years
- Mean age = 23.5 years
Table 2 shows a vast improvement between the first and second data collections, for delayed/no prescriptions in AIC, with 78.95% of patients received a delayed prescription or no prescription during the second collection, only just falling short of our standard.
|
|
Conclusions |
|---|
|
TOPNotesAbstractCriteria chosenStandard setPreparation and planningData collection 1Changes to be evaluatedData collection 2ConclusionsReferences |
|---|
We held a practice meeting on 30 March 2007 to discuss the audit results. The general expectation was of poor performance and everyone was pleasantly surprised to see the leap from less than 9% delayed prescribing to nearly 80%.
We discussed those four patients receiving immediate antibiotics, to see if any issues needed addressing:
- Treated by a locum doctor
- Symptoms documented for 4 days. You could argue that it was pointless waiting another day to treat
- ‘Recurrent conjunctivitis’ documented in the notes, however, they had only had two previous episodes over a 2-year period and the exclusion criterion was ‘chronic infection’
- Limited documentation making reason for immediate prescribing unclear
It was agreed that these cases did not raise any major difficulties around our management of AIC.
Some practical concerns were raised at the discussion.
- Where to keep prescriptions issued at the branch surgery, not least because the branch surgery is closed two afternoons per week. It was agreed that choice should be left to the patient, with a reminder about surgery opening times.
- Problems printing the PIL at the branch surgery. The computer system has since been reconfigured to allow printing of the PIL in the branch surgery.
- Practicalities of supplying PILs after telephone consultations. A difficult problem to solve, but it was agreed that patients should be offered the option to collect a PIL from reception if they wish.
- One particular nursery in our area consistently refused to allow children and staff to attend without a prescription for eye drops. I am planning to visit them in the near future for a discussion about their concerns, to see if a happy medium can be reached between us.
- Confusion over what to do with prescriptions at the weekends, when patients would be unable to collect prescriptions. In fact, most people found it easier to give the printed script to take away and documented this action.
The overall opinion was that the audit had been a positive experience and we were all encouraged by the results. The initial fears that this new management approach would lengthen previously straightforward consultations were not found to be the case by anyone. We had all found patient education easy and were pleasantly surprised at the receptiveness of our patients, many of whom were relieved not to need antibiotics. Even the most cynical GPs reported it to be a great learning experience.
There was still some concern about whether all this hard work had been justified. The following information helped allay those fears:
Delayed prescription collections
- Total number delayed scripts documented = 15
- Unknown outcome = 3 (scripts given to take away, two were before Christmas weekend, one was a T/R leaving area on same day)
- Scripts collected at later date = 4/12 (33%)
- Scripts not collected and no further contact for similar complaint = 8/12 (66%)
I had withheld the data collection for 3 weeks after the end of the audit cycle, to allow time for scripts to be collected and problems to be reported by patients. Our delayed prescription collection rate was lower than quoted research rates (Little et al., 1997, 2001; Dowell et al., 2001; Arroll et al., 2002, 2003; Cates 2003) and we were all encouraged by this.
Table 4 clearly demonstrates that during the first data collection, when we were automatically prescribing for AIC, appointments were given to nearly 50% of patients. During the second data collection, when we were using delayed prescriptions, nearly three-fourth of patients were managed via telephone consultation alone. We have already started to save ourselves appointment time and as our patients become more educated about the minor, self-limiting nature of AIC; we hope to further reduce those numbers of appointments.
|
Overall, I have found this to be a very satisfying audit. The topic was relevant to my everyday practice, it was a steep learning experience for everyone involved, it was easy to set up and there was great potential to change practice. More importantly, it has changed practice within our small community. Of course, those cynical GPs within our practice may well revert back to their old ways, now that they are no longer being audited. It is always more tempting to go for the ‘easy option’ and run off a quick script for antibiotics, especially when surgeries are running late and a mother is pleading for a script so her kids can go to school, but I would hope that the ease with which the changes have been instituted, the potential decrease in future workload and the robust new system that has been put in place to enable delayed prescribing within the practice would encourage them to continue with this new approach. There is significant potential to extend delayed prescribing to other minor, self-limiting conditions, thereby further reducing workload and risk of antibiotic resistance. For this reason, I would encourage repeated audit of the practice's management of AIC in a few years time, but would perhaps suggest extending the audit to include management of all minor self-limiting illnesses.
|
Notes |
|---|
|
TOPNotesAbstractCriteria chosenStandard setPreparation and planningData collection 1Changes to be evaluatedData collection 2ConclusionsReferences |
|---|
The Information Leaflet given to patients or carers (Appendix 1) and letter for patients (Appendix 2) used in this audit is available as supplementary data online
|
References |
|---|
|
TOPNotesAbstractCriteria chosenStandard setPreparation and planningData collection 1Changes to be evaluatedData collection 2ConclusionsReferences |
|---|
-
Arroll B, Kenealy T, Goodyear-Smith F, Kerse N. Delayed prescriptions can reduce antibiotic use in acute respiratory infections. British Medical Journal (2003) 327:1361–62.
Arroll B, Kenealy T, Kerse N. Do delayed prescriptions reduce the use of antibiotics for the common cold? A single-blind controlled trial. The Journal of Family Practice (2002) 51:324–28.[Web of Science][Medline]
Cates C. Delayed prescriptions in primary care. The British Journal of General Practice (2003) 836–37.
Department of Health. Prescription cost analysis—England (2005) Accessed via www.doh.gov.uk/publication and statistics [date last accessed 10.10.2006].
Dowell J, Pitkethly M, Bain J. A randomised controlled trial of delayed antibiotic prescribing as a strategy for managing uncomplicated respiratory tract infection in primary care. The British Journal of General Practice (2001) 51:200–05.[Medline]
Everitt H, Kumar S, Little P. A qualitative study of patients’ perceptions of acute infective conjunctivitis. The British Journal of General Practice (2003) 36–41.
Everitt H, Little P. How do GPs diagnose and manage acute infective conjunctivitis? A GP survey. Family Practice (2002) 19:658–60.
Everitt H, Little P, Smith P. A randomised controlled trial of management strategies for acute infective conjunctivitis in general practice. British Medical Journal (2006) 333:321.
Little P, Gould C, Williamson I, et al. Pragmatic randomised controlled trial of two prescribing strategies for childhood acute otitis media. British Medical Journal (2001) 322:336–42.
Little P, Williamson I, Warner G, Gould C, Gantley M, Kinmonth A. Open randomised trial of prescribing strategies in managing sore throat. British Medical Journal (1997) 314:722–27.
Prodigy. PIL: infective conjunctivitis. Updated October 2006.
Rietveld R, ter Riet G, Bindels P, Sloos J, van Weert H. Predicting bacterial cause in infectious conjunctivitis—cohort study on informativeness of combinations of signs and symptoms. British Medical Journal (2004) 329:206–10.
Rose P, Harnden A, Brueggemann A, Perera R, Sheikh A, Crook D, Mant D. Chloramphenicol treatment for acute infective conjunctivitis in children in primary care—a randomised double-blind placebo-controlled trial. The Lancet (2005) 366:9479.
Rose P, Ziebland S, Harnden A, Mayon-White R, Mant D. Why do general practitioners prescribe antibiotics for acute infective conjunctivitis in children? Qualitative interviews with GPs and a questionnaire survey or parents and teachers. Family Practice (2006) 23:226–32.
Sheikh A, Hurwitz B. Antibiotics versus placebo for acute bacterial conjunctivitis. Cochrane Database of Systematic Reviews (2006) [Search date 19.09.2006].
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||