False Positive Referrals Research Articles

Should tumbling E go out of date in amblyopia screening? Evidence from a population-based sample normative in children aged 3–4 years

AimsTo determine a normative of tumbling E optotype and its feasibility for visual acuity (VA) assessment in children aged 3-4 years.MethodsA cross-sectional study of 1756 children who were invited to...

ICD-11 Gaming Disorder: Needed and just in time or dangerous andmuchtooearly?

In their debate contribution, Aarseth et al. (2016) strongly argue against the proposal of WHO ICD-11 (International Classification of Diseases, 11th revision) to include Gaming Disorder as a new diagnostic category emphasizing the fact that no consensus exists on the definition and the risk that gaming will be demonized and gamers stigmatized resulting in a tsunami of false positive referrals to treatment. In this commentary, it is argued that gaming is indeed just another relatively innocent recreational activity with only a small minority losing control resulting in gaming-related problems. It is also argued that – despite a lack of full consensus on the diagnostic criteria – there are clear indications that Gaming Disorder is a relevant clinical entity worldwide and that official recognition as a mental disorder is urgently needed to facilitate the further development, accessibility, and reimbursement of the treatment.

Factors influencing accuracy of referral and the likelihood of false positive referral by optometrists in Bradford, United Kingdom

AimsLevels of false positive referral to ophthalmology departments can be high. This study aimed to evaluate commonality between false positive referrals in order to find the factors which may influence referral accuracy. MethodsIn 2007/08, a sample of 431 new Ophthalmology referrals from the catchment area of Bradford Royal Infirmary were retrospectively analysed. ResultsThe proportion of false positive referrals generated by optometrists decreases with experience at a rate of 6.2% per year since registration (p<0.0001). Community services which involved further investigation done by the optometrist before directly referring to the hospital were 2.7 times less likely to refer false positively than other referral formats (p=0.007). Male optometrists were about half as likely to generate a false positive referral than females (OR=0.51, p=0.008) and as multiple/corporate practices in the Bradford area employ less experienced and more female staff, independent practices generate about half the number of false positive referrals (OR=0.52, p=0.005). ConclusionsClinician experience has the greatest effect on referral accuracy although there is also a significant effect of gender with women tending to refer more false positives. This may be due to a different approach to patient care and possibly a greater sensitivity to litigation. The improved accuracy of community services (which often refer directly after further investigation) supports further growth of these schemes.

Preschool children's vision screening in New Zealand: a retrospective evaluation of referral accuracy

ObjectivesTo assess the accuracy of preschool vision screening in a large, ethnically diverse, urban population in South Auckland, New Zealand.DesignRetrospective longitudinal study.MethodsB4 School Check vision screening records (n=5572) were compared...

How accurate are optometrist referrals for glaucoma in the NICE era?

PurposeTo determine if the reason(s) stated by the optometrist affect the positive predictive value (PPV) of referrals for suspected glaucoma.MethodsOptometrist referrals in Portsmouth are streamlined though the Glaucoma Referral Refinement Scheme (GRRS). We retrospectively analysed 100 ‘positive’ and 100 ‘negative’ referrals to the GRRS. Positives: those referred on to a secondary care glaucoma clinic; negatives: those discharged from the GRRS. Reason(s) for referral were determined and classified: intraocular pressure (IOP) optic disc changes (OD), visual field defect (VF), shallow anterior chamber depth (ACD). Since negative referrals to GRRS are four times more prevalent than positives, we calculated a ‘corrected’ PPV for each reason by multiplying the number of negatives by 4.ResultsThe overall PPV was 0.16. 70.4% cases were referred for a single reason (47.2% IOP, 12.2% OD, 8.8% VF, 2.4% ACD); the PPV was 0.11 for IOP and for OD, 0.05 for VF, 0.17 for ACD. 26.2% were referred for two reasons; the PPV was 0.47 for IOP + OD, 0.21 for OD + VF and 0.14 for IOP + VF. 4.2% were referred for IOP + OD + VF with a PPV of 0.62. Positive family history of glaucoma (21.4% referrals) did not increase PPV. The GRRS increased the PPV to 0.8.ConclusionsThe majority of patients are referred for a single reason, yet the chance of a positive diagnosis is low. Accuracy is increased when IOP is reported in combination with OD, but VF adds little value as a second parameter. The highest PPV is found when all three reasons are stated (IOP + OD + VF). The proportion of referrals for IOP alone is similar to that found prior to the NICE guidelines. However, the PPV is considerably lower; 0.11 compared to 0.3–0.4 found in previous studies. In the NICE era, refinement schemes play an important role in reducing false positive referrals.

How accurate are optometrist referrals for glaucoma in the NICE era?

PurposeTo determine if the reason(s) stated by the optometrist affect the positive predictive value (PPV) of referrals for suspected glaucoma.MethodsOptometrist referrals in Portsmouth are streamlined though the Glaucoma Referral Refinement Scheme (GRRS). We retrospectively analysed 100 ‘positive’ and 100 ‘negative’ referrals to the GRRS. Positives: those referred on to a secondary care glaucoma clinic; negatives: those discharged from the GRRS. Reason(s) for referral were determined and classified: intraocular pressure (IOP) optic disc changes (OD), visual field defect (VF), shallow anterior chamber depth (ACD). Since negative referrals to GRRS are four times more prevalent than positives, we calculated a ‘corrected’ PPV for each reason by multiplying the number of negatives by 4.ResultsThe overall PPV was 0.16. 70.4% cases were referred for a single reason (47.2% IOP, 12.2% OD, 8.8% VF, 2.4% ACD); the PPV was 0.11 for IOP and for OD, 0.05 for VF, 0.17 for ACD. 26.2% were referred for two reasons; the PPV was 0.47 for IOP + OD, 0.21 for OD + VF and 0.14 for IOP + VF. 4.2% were referred for IOP + OD + VF with a PPV of 0.62. Positive family history of glaucoma (21.4% referrals) did not increase PPV. The GRRS increased the PPV to 0.8.ConclusionsThe majority of patients are referred for a single reason, yet the chance of a positive diagnosis is low. Accuracy is increased when IOP is reported in combination with OD, but VF adds little value as a second parameter. The highest PPV is found when all three reasons are stated (IOP + OD + VF). The proportion of referrals for IOP alone is similar to that found prior to the NICE guidelines. However, the PPV is considerably lower; 0.11 compared to 0.3–0.4 found in previous studies. In the NICE era, refinement schemes play an important role in reducing false positive referrals.

Glaucoma-service provision in Scotland: introduction and need for Scottish Intercollegiate Guidelines Network guidelines.

PurposeTo describe the pattern of glaucoma-service delivery in Scotland and identify areas for improvement, taking into account Scottish General Ophthalmic Services (GOS) arrangements and the Eye Care Integration project, and to design Scottish Intercollegiate Guidelines Network (SIGN) guidelines to refine the primary and secondary interface of glaucoma care.Materials and methodsA glaucoma-survey questionnaire was sent to all consultant glaucomatologists in Scotland. The design of SIGN guidelines was based on the results of the questionnaire using SIGN methodology.ResultsOver 90% of Scottish glaucoma care is triaged and delivered within hospital services. Despite GOS referral, information is variable. There are no consistent discharge practices to the community. These results led to defined research questions that were answered, thus formulating the content of the SIGN guidelines. The guideline covers the assessment of patients in primary care, referral criteria to hospital, discharge criteria from hospital to community, and monitoring of patients at risk of glaucoma.ConclusionWith increasing age and limitations to hospital resources, refining glaucoma pathways between primary and secondary care has become a necessity. Scotland has unique eye care arrangements with both the GOS and Eye Care Integration project. It is hoped that implementation of SIGN guidelines will identify glaucoma at the earliest opportunity and reduce the rate of false-positive referrals to hospital.

Vision screening – referral to discharge. Outcomes from a routine vision screening programme

Aims: To investigate a local vision screening pro- gramme, the conditions identified at routine vision screening and the visual acuity outcomes of children referred. Methods: A retrospective analysis of children who underwent vision screening between 1 September 2005 and 31 August 2006 was undertaken. The screening programme included an assessment of uniocular vision, cover test, ocular motility, binocular reflex and stereo-acuity. All children referred had their hospital case notes reviewed and data on final corrected visual acuity, refractive error and follow- up period collected. Results: Of 2468 children offered vision screening 2240 gave consent and were tested (90.8% coverage), 309 (13.8%) children were referred, and 264 (85.4%) patients attended of whom 33 (12.5%) were false positive referrals. The vision screening programme had a positive predictive value of 87.5%. Corrected visual acuity was 0.200 logMAR or better in each eye in 89.1% of patients, 64.0% required only optimum refractive correction as their sole treatment and 10.2% required a period of occlusion therapy. Conclusions: Orthoptic-based screening programmes provide an efficient vision screening mechanism, achieving high coverage, and low re-test and false positive levels. Vision screening has been shown to identify a variety of conditions which had until that point gone unnoticed. Those referred show excellent visual acuity outcomes.

Reassessing critical maternal antibody threshold in RhD alloimmunization: a 16-year retrospective cohort study

To determine the critical maternal antibody threshold for specialist referral in cases of RhD alloimmunization. This was a retrospective cohort study, covering a 16-year period at the national tertiary fetal medicine center for management of alloimmunization. Data concerning RhD alloimmunized pregnant women were extracted from an institutional database and maternal anti-D antibody levels were cross-checked with the national reference laboratory. Fetal hemoglobin (Hb) levels were determined only at the first intrauterine transfusion (IUT) and were compared with the pretransfusion maternal anti-D antibody level (IU/mL). Sensitivity, specificity and positive and negative predictive values of maternal antibody thresholds for detecting moderate to severe (Hb ≤ 0.64MoM) fetal anemia were calculated. Between 1996 and 2011, 66 women underwent a first IUT for RhD alloimmunization at our institution. The highest serum anti-D antibody level was extracted for 208 RhD alloimmunized women who did not require IUT during the last 10 years of the study period. The traditional maternal antibody threshold of > 15 IU/mL failed to detect 20% of cases of moderate to severe fetal anemia. The ≥ 4 IU/mL threshold had 100% sensitivity but a 45% false-positive rate. The optimal anti-D antibody threshold for specialist referral in our population was ≥ 6 IU/mL; at this level, no case of moderate to severe anemia was missed and specificity was 61%. Use of this threshold would have eliminated 10% of referrals to our fetal medicine unit without compromising fetal outcomes. Setting the critical maternal RhD antibody level at >15 IU/mL does not provide sufficient sensitivity. The lower threshold of ≥ 4 IU/mL, though sensitive, is associated with a 45% false-positive rate. In our population, a threshold of ≥ 6 IU/mL minimizes false-positive referrals while maintaining 100% sensitivity for moderate to severe fetal anemia.

Cambridge community Optometry Glaucoma Scheme

With a higher life expectancy, there is an increased demand for hospital glaucoma services in the United Kingdom. The Cambridge community Optometry Glaucoma Scheme (COGS) was initiated in 2010, where new referrals for suspected glaucoma are evaluated by community optometrists with a special interest in glaucoma, with virtual electronic review and validation by a consultant ophthalmologist with special interest in glaucoma. 1733 patients were evaluated by this scheme between 2010 and 2013. Clinical assessment is performed by the optometrist at a remote site. Goldmann applanation tonometry, pachymetry, monoscopic colour optic disc photographs and automated Humphrey visual field testing are performed. A clinical decision is made as to whether a patient has glaucoma or is a suspect, and referred on or discharged as a false positive referral. The clinical findings, optic disc photographs and visual field test results are transmitted electronically for virtual review by a consultant ophthalmologist. The number of false positive referrals from initial referral into the scheme. Of the patients, 46.6% were discharged at assessment and a further 5.7% were discharged following virtual review. Of the patients initially discharged, 2.8% were recalled following virtual review. Following assessment at the hospital, a further 10.5% were discharged after a single visit. The COGS community-based glaucoma screening programme is a safe and effective way of evaluating glaucoma referrals in the community and reducing false-positive referrals for glaucoma into the hospital system.

Diagnostic Pure-Tone Audiometry in Schools: Mobile Testing without a Sound-Treated Environment

To validate diagnostic pure-tone audiometry in schools without a sound-treated environment using an audiometer that incorporates insert earphones covered by circumaural earcups and real-time environmental noise monitoring. A within-subject repeated measures design was employed to compare air (250 to 8000 Hz) and bone (250 to 4000 Hz) conduction pure-tone thresholds measured in natural school environments with thresholds measured in a sound-treated booth. 149 children (54% female) with an average age of 6.9 yr (SD = 0.6; range = 5-8). Average difference between the booth and natural environment thresholds was 0.0 dB (SD = 3.6) for air conduction and 0.1 dB (SD = 3.1) for bone conduction. Average absolute difference between the booth and natural environment was 2.1 dB (SD = 2.9) for air conduction and 1.6 dB (SD = 2.7) for bone conduction. Almost all air- (96%) and bone-conduction (97%) threshold comparisons between the natural and booth test environments were within 0 to 5 dB. No statistically significant differences between thresholds recorded in the natural and booth environments for air- and bone-conduction audiometry were found (p > 0.01). Diagnostic air- and bone-conduction audiometry in schools, without a sound-treated room, is possible with sufficient earphone attenuation and real-time monitoring of environmental noise. Audiological diagnosis on-site for school screening may address concerns of false-positive referrals and poor follow-up compliance and allow for direct referral to audiological and/or medical intervention.

Results of a community vision-screening program using the Spot photoscreener

To describe the results of vision screenings performed with the Spot photoscreener in the community setting. Low-income, predominantly Hispanic children in day care and preschool settings were screened by lay operators using the Spot photoscreener. Inclusion criteria were age 6-72 months and availability of a complete photoscreening record. Referral criteria were based on Vision Screening Committee of American Association for Pediatric Ophthalmology and Strabismus guidelines. Data were stratified by age group and analyzed for percentage of children referred for hyperopia, myopia, astigmatism, anisometropia, anisocoria, and ocular misalignment. Vision screening records were compared with comprehensive eye examination records from an optometrist or ophthalmologist to determine positive predictive value. Vision screening examinations were performed on 8,317 subjects from September 2011 through May 2012. The mean age of the 7,814 subjects (3953 males) meeting inclusion criteria was 44.4 months. The Spot referred 2,393 (30.6%). Of the screened population, the suspected reason for referral was astigmatism in 1,863 (23.8%), ocular misalignment in 879 (11.3%), anisometropia in 90 (1.2%), myopia in 82 (1.1%), hyperopia in 63 (0.8%), and anisocoria in 16 (0.2%). Comprehensive examination reports, including a cycloplegic refraction, were available for 300 referred children (12.5%). The reason for referral was confirmed in 55.7%, with an overall positive predictive value of 65.7%. The Spot photoscreener yielded a high overall referral rate. Although a high prevalence of astigmatism may be expected in this population, a high referral rate for suspected ocular misalignment led to a very high proportion of false positive referrals, suggesting that the software for this algorithm is in need of refinement.

Re-attendance after false-positive screening mammography: a population-based study in the Netherlands

Background:In the current study, mammography adherence of women who had experienced a false-positive referral is evaluated, with emphasis on the probability of receiving surveillance mammography outside the national screening programme.Methods:We included 424 703 consecutive screens and collected imaging, biopsy and surgery reports of 3463 women who experienced a false-positive referral. Adherence to screening, both in and outside the screening programme, was evaluated.Results:Two years after the false-positive referral, overall screening adherence was 94.6%, with 64.7% of women returning to the national screening programme, compared with 94.9% of women re-attending the screening programme after a negative screen (P<0.0001). Four years after the false-positive screen, the overall adherence had decreased to 85.2% (P<0.0001) with a similar proportion of the women re-attending the screening programme (64.4%) and a lower proportion (20.8%) having clinical surveillance mammography. Women who had experienced a false-positive screen at their first screening round were less likely to adhere to mammography than women with an abnormal finding at one of the following screening rounds (92.4% vs 95.5%, P<0.0001).Conclusion:Overall screening adherence after previous false-positive referral was comparable to the re-attendance rate of women with a negative screen at 2-year follow-up. Overall adherence decreased 4 years after previous false-positive referral from 94.6% to 85.2%, with a relatively high estimate of women who continue with clinical surveillance mammography (20.8%). Women with false-positive screens should be made aware of the importance to re-attend future screening rounds, as a way to improve the effectiveness of the screening programme.

Optimisation of volume-doubling time cut-off for fast-growing lung nodules in CT lung cancer screening reduces false-positive referrals

Optimisation of volume-doubling time cut-off for fast-growing lung nodules in CT lung cancer screening reduces false-positive referrals

Multimodal hyperspectroscopy as a triage test for cervical neoplasia: Pivotal clinical trial results

ObjectiveTo prospectively evaluate a new non invasive device that combines fluorescence and reflectance spectroscopy in a population in women at risk for cervical dysplasia. MethodsA total of 1607 women were evaluated with multimodal hyperspectroscopy (MHS), a painless test with extremely high spectral resolution. Subjects who were referred to colposcopy based on abnormal screening tests or other referral criteria underwent the MHS test and also had a sample taken for additional cytology and presence of high risk human papilloma virus (HPV) prior to undergoing biopsy. ResultsSensitivity of MHS for cervical intraepithelial neoplasia (CIN) 2+ was 91.3% (252/276). Specificity, or the potential reduction in referrals to colposcopy and biopsy, was 38.9% (222/570) for women with normal or benign histology and 30.3% (182/601) for women with CIN1 histology. Two year follow-up data, collected for a subgroup of 804 women, revealed 67 interval CIN2+ that originally were diagnosed at enrollment as normal or CIN1. MHS identified 60 of these (89.6%) as positive for CIN2+ prior to their discovery during the two year follow-up period. ConclusionsMHS provides an immediate result at the point of care. Recently, the limitations of cytology have become more obvious and as a consequence greater emphasis is being placed on HPV testing for cervical cancer screening, creating a need for an inexpensive, convenient and accurate test to reduce false positive referrals to colposcopy and increase the yield of CIN2+ at biopsy. MHS appears to have many of the attributes necessary for such an application.

Optimisation of volume-doubling time cutoff for fast-growing lung nodules in CT lung cancer screening reduces false-positive referrals

To retrospectively investigate whether optimisation of volume-doubling time (VDT) cutoff for fast-growing nodules in lung cancer screening can reduce false-positive referrals. Screening participants of the NELSON study underwent low-dose CT. For indeterminate nodules (volume 50-500 mm(3)), follow-up CT was performed 3 months after baseline. A negative baseline screen resulted in a regular second-round examination 1 year later. Subjects referred to a pulmonologist because of a fast-growing (VDT <400 days) solid nodule in the baseline or regular second round were included in this study. Histology was the reference for diagnosis, or stability on subsequent CTs, confirming benignity. Mean follow-up of non-resected nodules was 4.4 years. Optimisation of the false-positive rate was evaluated at maintained sensitivity for lung cancer diagnosis with VDT <400 days as reference. Sixty-eight fast-growing nodules were included; 40 % were malignant. The optimal VDT cutoff for the 3-month follow-up CT after baseline was 232 days. This cutoff reduced false-positive referrals by 33 % (20 versus 30). For the regular second round, VDTs varied more among malignant nodules, precluding lowering of the VDT cutoff of 400 days. All malignant fast-growing lung nodules referred after the 3-month follow-up CT in the baseline lung cancer screening round had VDT ≤232 days. Lowering the VDT cutoff may reduce false-positive referrals. • Lung nodules are common in CT lung cancer screening, most being benign • Short-term follow-up CT can identify fast-growing intermediate-size lung nodules • Most fast-growing nodules on short-term follow-up CT still prove to be benign • A new volume-doubling time (VDT) cut-off is proposed for lung screening • The optimised VDT cutoff may decrease false-positive case referrals for lung cancer.

There is a trend for the diagnosis of glaucoma to be made at an earlier stage in 2010 compared to 2008 in Oxford, United Kingdom

New guidelines on glaucoma diagnosis and management were published by the National Institute of Clinical Excellence (NICE) in 2009. The aim of this study was to determine whether these guidelines on glaucoma referral have changed the numbers of patients diagnosed with glaucoma in Oxford and whether the severity of disease at presentation has altered between 2008 and 2010. A retrospective study was undertaken of patients referred to a glaucoma case-finding clinic at the Oxford Eye Hospital in 2008 and 2010. Severity of glaucoma was assigned on the basis of the worse eye and defined according to mean deviation (MD) on 24-2 Humphrey visual field testing, with early glaucomatous visual field defects defined as MD less than -6 dB, moderate defects as MD between -6 and -12 dB and severe defects as MD more than -12 dB. Statistical analysis was undertaken using Chi squared and Mann Whitney tests. The number of referrals to this clinic increased from 521 patients in 2008, to 895 patients in 2010. 85 (16.3%) patients were diagnosed with ocular hypertension in 2008, compared to 144 (16.1%) in 2010. The number of patients diagnosed with glaucoma increased from 94 cases in 2008 to 113 cases in 2010, although the percentage of referrals diagnosed with glaucoma fell from 18.1% to 12.6%. More patients diagnosed with glaucoma in 2010 had early disease (61.9% in 2010 vs 48.8% in 2008, p = 0.03). Patients found to have glaucoma in 2010 had less severe visual field defects than in 2008: patients presented with a lower mean deviation (-6.36 dB in 2010, -7.95 dB in 2008, p = 0.03) and lower pattern standard deviation (5.44 dB in 2010, 6.64 dB in 2008, p = 0.02). However, there was no significant difference between age of presentation (65.1 years in 2010, 65.2 years in 2008, p = 0.82), and IOP (21.62 mmHg in 2010, 23.15 mmHg in 2008, p = 0.22). There has been a significant increase in number of referrals to the glaucoma case-finding clinic in Oxford between 2008 and 2010. The introduction of NICE referral guidelines in 2009 has had the benefit of increasing the absolute number of patients detected with glaucoma and has resulted in more patients being diagnosed with early disease. More effort needs to be made to reduce the number of false positive referrals. In addition, the guidelines on referral of individuals with ocular hypertension need to be reconsidered.

Trends in outpatient follow-up rates in England

Trends in the follow-up rate for a variety of outpatient specialties are investigated over a 24 year period. In the majority of specialties the greatest reduction in the follow-up rate occurred prior to 2000/01. Many specialties exhibit a degree of cyclic behaviour in the follow-up rate which seems to arise out of non-linear behaviour over time in the rates of GP referral. Sudden increases in GP referral can arise from a congruence of adverse environmental conditions; i.e. sporadic ‘spike’ events, or due to longer term cycles in the wider environment (including infectious outbreaks). Such increases result in a higher proportion of false positive GP referrals leading to a dip in the follow-up rate. GP commissioners in England are advised to pay far greater attention to the undulating nature of outpatient activity and costs. Potential environmental factors contributing to such cycles are discussed along with the implications to activity and cost forecasting.

Comparing the Diagnostic Yields of Technologists and Radiologists in an Invitational Colorectal Cancer Screening Program Performed with CT Colonography

To compare the diagnostic yields of a radiologist and trained technologists in the detection of advanced neoplasia within a population-based computed tomographic (CT) colonography screening program. Ethical approval was obtained from the Dutch Health Council, and written informed consent was obtained from all participants. Nine hundred eighty-two participants (507 men, 475 women) underwent low-dose CT colonography after noncathartic bowel preparation (iodine tagging) between July 13, 2009, and January 21, 2011. Each scan was evaluated by one of three experienced radiologists (≥800 examinations) by using primary two-dimensional (2D) reading followed by secondary computer-aided detection (CAD) and by two of four trained technologists (≥200 examinations, with colonoscopic verification) by using primary 2D reading followed by three-dimensional analysis and CAD. Immediate colonoscopy was recommended for participants with lesions measuring at least 10 mm, and surveillance was recommended for participants with lesions measuring 6-9 mm. Consensus between technologists was achieved in case of discordant recommendations. Detection of advanced neoplasia (classified by a pathologist) was defined as a true-positive (TP) finding. Relative TP and false-positive (FP) fractions were calculated along with 95% confidence intervals (CIs). Overall, 96 of the 982 participants were referred for colonoscopy and 104 were scheduled for surveillance. Sixty of 84 participants (71%) referred for colonoscopy by the radiologist had advanced neoplasia, compared with 55 of 64 participants (86%) referred by two technologists. Both the radiologist and technologists detected all colorectal cancers (n = 5). The relative TP fraction (for technologists vs radiologist) for advanced neoplasia was 0.92 (95% CI: 0.78, 1.07), and the relative FP fraction was 0.38 (95% CI: 0.21, 0.67). Two technologists serving as a primary reader of CT colonographic images can achieve a comparable sensitivity to that of a radiologist for the detection of advanced neoplasia, with far fewer FP referrals for colonoscopy.

Review of preschool vision screening referrals in South Auckland, New Zealand

Background: Limited data are available on the causes of visual impairment in preschool children in New Zealand. We aimed to review demographic and visual parameters in children referred to the Ophthalmology Department, Manukau Super Clinic from vision screening programs in South Auckland.Methods: Retrospective medical record review of 131 children, aged three to five years, referred from community‐based vision screening programs to the Ophthalmology Department for further assessment. Medical records were reviewed to determine: the reason for referral; findings from ophthalmic assessments; treatment received; and visual acuity at the final visit. The main outcome measures were the cause of visual impairment in children referred from preschool vision screening and the visual acuity at the final follow‐up visit.Results: Thirty‐eight (29.0 per cent) children were discharged after their initial assessment as false positive referrals. Almost half (45.5 per cent) of the children were prescribed glasses for the correction of refractive error, amblyopia or strabismus. Twenty‐nine (22.1 per cent) children were diagnosed with amblyopia with an average follow‐up period of 17.5 ± 2.7-months. In general, compliance with therapy for amblyopia was poor with 48.3 per cent non‐compliant with their prescribed treatment regimen. Despite this, visual outcomes were good with an average final visual acuity in the amblyopic eye of 0.294 ± 0.231-logMAR (Snellen 6/12).Conclusions: The ‘positive predictive value’ for the Counties‐Manukau preschool vision screening program was 47.4 per cent, suggesting that the visual acuity measurements alone produce a significant number of false positive results. In children diagnosed with amblyopia, early detection and intervention showed significant improvement in vision in the amblyopic eye, with many children also showing improved binocular function.