A review of 1000 referrals to Walsall s hospital eye service

Transcription

1 Journal of Public Health Vol. 38, No. 3, pp doi: /pubmed/fdv081 Advance Access Publication June 14, 2015 A review of 1000 referrals to Walsall s hospital eye service M. Fung 1,P.Myers 1, P. Wasala 1, N. Hirji 2 1 Public Health Department, Walsall Council, Walsall WS1 1TP, UK 2 Academic Unit of Public Health, University of Leeds, Leeds LS2 9JT, UK Address correspondence to Matthew Fung, matthew.fung@nhs.net ABSTRACT Background Referrals to ophthalmology are predominantly made from general practitioners (GPs) and optometrists. These two groups of referrers receive differing types and levels of training and are equipped with different instrumentation. The purpose of this study was to determine whether the quality of referrals to the hospital eye service (HES) differs between GPs and optometrists in Walsall. Methods Referrals into the HES were identified from Q retrospectively until 1000 notes had been reached. Each record was scrutinized using a standard template. Data were analysed and summary statistics produced including positive predictive values and interobserver agreement. Results We achieved our target of auditing 1000 records. The false-positive rate (patients being discharged from HES with a normal vision diagnosis) was 7.7% of referrals from GPs and 6.2% of referrals from optometrists. Concordance between referred condition and diagnosed condition at HES between optometrists and ophthalmologists was 76.1%, and between GPs and ophthalmologists was 67.2%. Conclusions In view of findings from this study, it is important for commissioners in the new reconfigured National Health Service to ensure that enhanced ophthalmic services are commissioned only on the basis of hard evidence sourced from local data rather than opinion or on data from another geographical area. Keywords eye disorders, primary care, secondary and tertiary services Introduction Referrals to ophthalmology are predominantly made from primary care: 43 49% from general practitioners (GPs), and 39 57% from primary care optometrists. 1 These two groups of referrers receive differing types and levels of training and are equipped with different instrumentation in their daily practice. Pierscionek et al. 2 suggested that there is likely to be variation in the types of cases and conditions seen by GPs and optometrists, but that there should not be a difference between the two groups in terms of accurate, high quality and efficient referrals. A personal view was published in the BMJ in March 2014 by Michael Clarke, a consultant ophthalmologist, which has certainly provoked debate. 3 The premise of this piece is that National Health Service (NHS) sight tests are essentially an unregulated, unsystematic screening programme, to which patients do not consent, and which have not been scrutinized by the National Screening Committee. Clarke acknowledges that some patients benefit from referral to detect early eye conditions, but states that there are many false-positive referrals (with associated unnecessary anxiety and societal costs). Prior to this audit, the accuracy, quality, type and outcome of referrals to the hospital eye service (HES) in Walsall was unknown, and we did not know the level of false-positive referrals to the local hospital. We were also unclear about whether systematic differences in referrals existed between optometrists and GPs. The purpose of this study was to explore any differences in referrals to the HES from GPs and optometrists in Walsall, particularly with regard to conditions referred, referral quality, concordance between referred diagnosis and ophthalmic diagnosis on discharge. We worked with the local hospital to M. Fung, Public Health Specialty Registrar P. Myers, Consultant in Public Health Medicine P. Wasala, Foundation Year 2 Doctor N. Hirji, Visiting Research Fellow # The Author Published by Oxford University Press on behalf of Faculty of Public Health. All rights reserved. For permissions, please journals.permissions@oup.com. 599

2 600 JOURNAL OF PUBLIC HEALTH conduct an audit of 1000 referrals to the ophthalmology department, identified through the hospital outpatient booking systems. Hand searching of each of these case notes was then undertaken. To our knowledge, this is the largest audit to date in the UK on referrals from primary care to the HES. Method We sought and gained approval from the Caldicott Guardian at Walsall Manor Hospital and from the Clinical Commissioning Group to undertake this audit. A report was run from the hospital outpatient database to identify referrals into the HES from Q retrospectively until 1000 notes containing a referral had been reached. We employed bank staff to locate and transport patient notes from record storage to the audit room. Processes and systems to ensure timely retrieval and couriering of notes were agreed with the hospital audit and compliance team. We developed a data collection template in Excel using automated fields, drop down lists and data validation mechanisms where possible to minimize recording error. The spreadsheet was structured to clearly record details of the referral, hospital activity and outcome of the referral. The presence of the following information from referral letters was recorded: date of referral; source of referral; address of referrer; patient name; date of birth; address; current treatment; visual acuity; intraocular pressure (IOP); visual fields; fundus comments/image; cup:disc ratio (C:D ratio); provisional diagnosis present (and specific details) and the degree of urgency. Hospital activity data included: date of first consultation; date of discharge; number of follow-up visits and ophthalmologist diagnosis. The following information from the outcome of each referral was recorded: letter sent to referring practitioner and concordance of referred and HES diagnoses. Patient identifiable data were not recorded, except the hospital number to enable retrieval of specific records to correct any identified data problems at a later point. A smaller subset of records recorded specifically whether patients had been referred with a cataract and whether the patient was subsequently listed for surgery or not. This question was pertinent as optometrists receive a payment for directly referring patients with cataract to ophthalmologists. Optometrists should discuss with patients prior to referring about their desire to have surgery, therefore the conversion rate from referral to surgery should be high. The Royal College of Ophthalmologists suggest that this system should be reviewed, and that the fee payable is only justified if the patient ultimately undergoes surgery. 4 Four members from Walsall Public Health audited records between April and July As some patient records contained multiple referral letters into the HES, we evaluated only the most recent HES referral letter. In running the original report from the hospital database, we were unable to distinguish at the outset between where the referral originated, and where within the HES the patient was being seen. Therefore, we expected to identify referrals predominantly from optometrists and GPs, and also from the orthoptic screening service, retinal screening and internal hospital referrals. We recorded data on all types of referral, but in this paper our focus is mostly on referrals from optometrists and GPs. Data were analysed predominantly in Excel using pivot tables and pivot graphs. Positive predictive values (PPVs) for optometrists and GPs were calculated based on concordance in referred and discharge diagnoses. Interobserver agreement comparing GPs and optometrists with ophthalmologists was calculated for selected conditions using the Kappa statistic, and further analysed using cross tabulated data. Results Sample of records We achieved our target of auditing 1000 patient records containing referrals into the HES. Of these, 569 referrals were made from optometrists, 143 referrals were from GPs and the remaining 288 were referrals from other sources, including the orthoptic screening service (161 referrals), retinal screening (70 referrals) or internal/other (57 referrals). A total of 906 records audited (90.6%) were from 2008 onwards, and 94 records (9.4%) were from 1991 to The distribution of records is left skewed (this is expected as we requested records from 2014 working retrospectively until 1000 records were reached). The full distribution by year of the referrals audited is shown in Fig. 1. Referral letter completeness One hundred per cent of referral letters from both GPs and optometrists contained basic information including patient name, address and address of the referrer. All records apart from one GP referral letter included date of birth. Optometric and medical-specific information contained within referral letters was heterogeneous (Fig. 2). As expected, inclusion of visual acuity, IOP, C:D ratio and fundus comments in optometrist referrals was far greater than GP referral letters. However, 101 GP referrals (71%) included details of patient history and medications, compared with 327 optometrist referrals (57%). A provisional diagnosis was recorded in 122 (85%) GP referrals and 527 (93%) optometrist referrals. Degree of urgency A marker of referral quality was the presence of degree of urgency within referral letters (Table 1). This was seldom

3 A REVIEW OF 1000 REFERRALS TO WALSALL S HOSPITAL EYE SERVICE Number of records Fig. 1 Patient records audited, by year of most recent referral all categories of referrer. Percentage 100.0% 90.0% 80.0% 70.0% 60.0% 50.0% 40.0% 30.0% 20.0% 10.0% 0.0% Patient name Patient address Address of referrer Patient DOB Fig. 2 Completeness of referral letters from GPs and optometrists. Visual acuity Year IOP Patient history/meds Referral letter item C:D ratio Fundus comments Provisional diagnosis GP Optometrist stated in GP and optometrist referrals with 666 (94%) referrals being devoid of any indication of urgency. The percentage of records achieving an appointment within the suggested referred urgency is also given (routine ¼ 12 weeks; soon ¼ 4 weeks; urgent ¼ 1 week). Time from referral to first appointment Analyses were performed on the number of days from any referral source to first appointment at the HES. A histogram illustrating this distribution is given in Fig. 3. A total of 638 (64%) of referrals were seen within 12 weeks (84 days), which

4 602 JOURNAL OF PUBLIC HEALTH Table 1 The presence of degree of urgency in GP and optometrist referrals to HES, and the percentage of these achieving referred degree of urgency Degree of urgency GP referrals Optometrist referrals % records containing degree of urgency (of which % referrals seen within suggested urgency) % records containing degree of urgency (of which % referrals seen within suggested urgency) Not detailed 79.72% 69.42% Routine within 12 weeks to HES 18.18% (68% seen within 12 weeks) 18.98% (69% seen within 12 weeks) Soon within 4 weeks to HES 1.40% (11% seen within 4 weeks) 7.21% (9% seen within 4 weeks) Urgent within 1 week to HES 0.70% (13% seen within 1 week) 4.39% (12% seen within 1 week) Total % % Number of records Fig. 3 Distribution of number of days from referral date to first consultation. is equivalent to a routine appointment, and 362 (36%) of referrals resulted in appointments outside of 12 weeks. Referred diagnosis 651 of 712 GP and optometrist referrals (91%) contained a referred diagnosis. The most commonly referred conditions from GPs and optometrists are shown in Fig. 4. Disorders of the lens (including cataract) was the most common reason for referral from optometrists (22% of all optometrist referrals). Disorders of eyelid, lacrimal system and orbit was the Number of days from referral to first HES appointment most common reason for referral from GPs (40% of all GP referrals). Cataract referrals A subset of data (170 referrals made by GPs or optometrists) was scrutinized for cataract referral details in addition to all other audit data. Of this subset, 30 referrals to the HES were specifically for cataract, which resulted in 24 referrals (80% of total) being listed. Of the 30 referrals for cataract, 26 referrals were made from optometrists and 4 referrals were made from

5 A REVIEW OF 1000 REFERRALS TO WALSALL S HOSPITAL EYE SERVICE 603 % total referred conditions for GPs/optometrists 45.00% 40.00% 35.00% 30.00% 25.00% 20.00% 15.00% 10.00% 5.00% 0.00% Fig. 4 Comparison of most commonly referred conditions by GP and optometrist. GPs. Five (19%) and 1 (25%) cataract referrals from optometrists and GPs, respectively, did not lead to patients being listed. False positives, PPVs The false-positive rate (i.e., patients being discharged from HES with a normal vision diagnosis) was 7.7% of referrals for GPs and 6.2% of referrals for optometrists. Concordance (PPV) in referred condition and diagnosed condition at HES between optometrists and ophthalmologists was 76.1%, and between GPs and ophthalmologists was 67.2%. The remaining referrals either did not have a referred diagnosis or did not concur with the diagnosis by the HES (23.9 and 32.8% for optometrists and GPs, respectively). These were sent by the referring practitioner due to raised levels of suspicion of something untoward and considered to require specialist intervention by the HES. Kappa statistic Two of the most common types of referral to the HES ( disorders of the lens and disorders of the eyelid/lacrimal system/ orbit ), and two less frequently referred conditions ( disorders of Disorders of lens IOP 21mm or greater Disorders of eyelid, lacrimal system and orbit Disorders of muscles, binocular movement/vision, amblyopia, accommodation and refraction Primary open angle glaucoma suspect Suspicious/abnormal disk Other disorders of retina Referred condition the cornea and diabetic retinopathy ) were analysed for interobserver agreement between the referrer and ophthalmologist using the kappa statistic (Table 2). 5 Cross tabulations of GP and optometrist referred diagnosis compared with ophthalmologist diagnosis for these four disorders are also given (Table 3). Disorders of the lens, one of the most commonly referred conditions, resulted in a kappa statistic of 0.91 (almost perfect agreement) for optometrists and 0.76 (substantial agreement) for GPs. The cross tabulated percentages when comparing referral for disorders of the lens against ophthalmological diagnosis were 98 and 100% for optometrists and GPs, respectively. Discussion Visual field defect Disorders of the cornea Optometrist GP Diabetic retinopathy Main findings of this study The main purpose of this study was to determine whether any differences exist between referrals from GPs and optometrists in Walsall. Referral letter completeness was found to be excellent for basic information provided by optometrists and GPs which was reassuring. However, we must be careful in interpreting the heterogeneity in optometric and medical-specific information provided between optometrists and GPs. The Disorders of choroid

6 604 JOURNAL OF PUBLIC HEALTH Table 2 Measure of interobserver agreement (kappa) between referring practitioner and ophthalmologist Disorder tested Number of referrals (Optometrist, GP) Optometrist/ ophthalmologist (kappa) Disorders of eyelid, 23, lacrimal system and orbit Disorders of the lens 125, (including cataract) Disorders of cornea 20, Diabetic retinopathy 39, GP/ ophthalmologist (kappa) Kappa is interpreted on a linear scale:,0 ¼ less than chance agreement; ¼ slight agreement; ¼ fair agreement; ¼ moderate agreement; ¼ substantial agreement; ¼ almost perfect agreement. differences of course may be a reflection of the conditions detected and referred in addition to the differences in training and equipment, for example a GP would be unlikely to provide fundus comments in referring a chalazion. The analysis of time from referral to first appointment yielded mixed results. Although 63.9% of referrals were seen within 12 weeks, around a third of referrals were seen beyond 12 weeks, a minority of which were seen 6 12 months after referral. We recognize that the majority of records did not contain any mention of degree of urgency, but we would consider 12 weeks to be a standard target to work towards for routine referrals. We also recognize that referrals taking many months to reach an appointment were often as a result of multiple did not attend episodes, and so this was largely out of the hospital s control. When comparing GP referrals with HES diagnosis, and optometrist referrals with HES diagnosis, PPVs were 67.2 and 76.1%, respectively. These PPVs are supported by kappa statistics for the four referred indications analysed, which found that the level of agreement between GPs and optometrists referring into the HES was at least substantial (and almost perfect agreement was found for 2/4 optometrist referrals and 1/4 GP referrals in the kappa analysis). What is already known on this topic? Davey et al. 1 previously found that the proportion of referrals to the HES from optometrists has been increasing over time, from 1988 to Various questions were raised around this observation, such as are GPs referring more patients to optometrists, and are patients choosing to attend optometrists instead of GPs? Pierscionek et al. 2 observed a larger number of referrals from optometrists compared with GPs. It was noted that the concurrence between referrer and ophthalmologist was different depending on the source of referral and condition referred. For instance, the concurrence between optometrists and ophthalmologists was higher for glaucoma referrals than for GPs, but the opposite was true for lid/tear duct/conjunctivitis referrals. Pooley and Frost 6 also conducted an audit of referrals which suggested that the added value of an optometrist referring patients via their GP to the HES is low. Lash 7 considered referral quality and suggested that content of referral forms should be improved and standardized, ideally to include specific information on cataract referrals. What this study adds Our data do not support any opinion that would suggest that referrals from optometrists following community-based sight tests result in disproportionally high levels of false-positives for our borough. The false-positive rates of only 7.7% of referrals from GPs and 6.2% of referrals from optometrists in this study were very similar to those observed by Pierscionek et al. We found good concordance in referred and ophthalmologist diagnosis which is very similar to other published studies. However, some patients referred into the HES will inevitably lead to a degree of false positives. Even though the percentage of normal vision diagnoses was low, primary care practitioners should still strive to refer carefully particularly given increasing service demand, budgetary pressures and constraints faced by the healthcare system. The efficacy of HES appointment reminder systems could be reviewed to help improve appointment attendance. The fact that some 85% of those referred urgently to the HES were not seen by the hospital within a week would suggest that perhaps it is time to encourage direct referrals from optometrists of all urgent referrals to avoid any delays caused by referrals going via the general medical practitioners. 8 Referral forms used by GPs and optometrists were heterogeneous, using very different structures and containing differing levels of information. A standard template should be considered and disseminated to improve referral quality further, particularly stressing the importance of record completeness and indicating urgency. Such a form does exist in Walsall and was disseminated to local optometrists but is not currently utilized universally nor does the General Ophthalmic Services contract oblige them to use this instead of the out dated GOS 18 form. Cataract referral forms used by optometrists are largely standardized, which may be one reason why we observed 80% of referrals for cataract resulting in being listed for surgery

7 A REVIEW OF 1000 REFERRALS TO WALSALL S HOSPITAL EYE SERVICE 605 Table 3 Referrals by GPs and optometrists compared with diagnosis by an ophthalmologist Opthalmologist diagnosis Percentage (%) referrals matching Total ophthalmologist diagnosis IOP 21 mm or greater Other Disorders of conjunctiva Normal vision Primary open angle glaucoma suspect Diabetic retinopathy Disorders of the cornea Disorders of lens Disorders of eyelid, lacrimal system and orbit GP reason for referral Disorders of eyelid, lacrimal system and orbit (79%) Disorders of lens (100%) Disorders of the cornea (85%) Diabetic retinopathy (80%) Optometrist reason for referral Disorders of eyelid, lacrimal system and orbit (86%) Disorders of lens (98%) Disorders of the cornea (80%) Diabetic retinopathy (92%) Rows refer to GP and optometrist reasons for referral, and columns refer to the diagnosis made by ophthalmology. Values highlighted in bold are where reason for referral concurred with diagnosis at ophthalmology. (notwithstanding the small sub-sample size). This represents a 33% higher conversion rate from referral for cataract to surgery than the rate previously reported by Lash. 7 In view of findings from this study, it is important for commissioners in the new reconfigured NHS to ensure that enhanced ophthalmic services are commissioned only on the basis of hard evidence sourced from local data rather than opinion or on data from another geographical area, even if the areas are considered to be similar. Limitations of this study This was the first audit with the local hospital in Walsall since the public health service relocated under jurisdiction of the Local Authority. We encountered a number of difficulties in gaining access to patient records, even when an audit proposal had been signed off by the Caldicott Guardian and the Clinical Commissioning Group. We are hopeful that future audits will be easier now relationships with the correct personnel have been established, and systems have been developed. This audit analysed 1000 records which, to our knowledge, is the largest audit of ophthalmology referrals in the UK. Our preference would have been to randomly sample patient referral case notes but due to difficulties in collating and gaining access to records, we were obliged to audit the identified and available referrals provided. This meant that some older records were identified and audited but were small in number (,10% of records pre-2008). We would of course aim to strengthen this aspect of our methodology for future audit work. Acknowledgements We acknowledge and thank Mrs Seeta Wakefield (Public Health Specialty Registrar) for helping to collect and record audit data. We also wish to thank Walsall Healthcare NHS Trust and Walsall Clinical Commissioning Group for their support and for providing access to records. Funding This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors. References 1 Davey CJ, Green C, Elliott DB. Assessment of referrals to the hospital eye service by optometrists and GPs in Bradford and Airedale. Ophthalmic Physiol Opt. 2011;31:23 8.

8 606 J O U R NA L O F P U B L IC H E A LT H 2 Pierscionek TJ, Moore JE, Pierscionek BK. Referrals to ophthalmology: optometric and general practice comparison. Ophthal Physiol Opt 2009;29: Clarke M. NHS sight tests include unevaluated screening examinations that lead to waste. BMJ 2014;348:g College statement on access to cataract surgery. The Royal College of Ophthalmologists. itemid=491 (October 2014, date last accessed). 5 Viera AJ, Garrett JM. Understanding interobserver agreement: the kappa statstic. Fam Med 2005;37(5): Pooley JE, Frost EC. Optometrists referrals to the hospital eye service. Ophthal Physiol Opt 1999;19(Suppl 1001):S Lash SC. Assessment of information included on the GOS 18 referral form used by optometrists. Ophthal Physiol Opt 2003;23: Menon GJ, Faridi UA, Gray RH. Direct referral of posterior capsular opacification by optometrists. Ophthal Physiol Opt 2004;24(2):