|
|
|
|
|
|
| SCREENING PROCEDURES IN OPHTHALMOLOGY | |
|
The discipline of Epidemiology encompasses a wide range of options varying from research methodology and estimation of risk at one end to more practical, pragmatic preventive strategies at the other end of the spectrum. One of these preventive strategies is screening, and this has found wide application the world over, in the recent past.
Screening attempts to classify populations as apparently normal or apparently abnormal. The abnormal individuals need to receive specified interventions after establishing the diagnosis, while the normal individuals need to be re-examined at periodic intervals to detect the future occurrence of the abnormality. The process of screening can be accomplished in different ways, by using different methods:
This can be classified as:
Screening and Surveys For far too long, there has been a tendency to use the two terms 'screening' and 'surveys' in a synonymous fashion. This is similar to the confusion that most clinicians have when they use the terms 'incidence' and 'prevalence'. It needs to be emphasized that screening and surveys are two completely different and independent epidemiologic tools. The major differences between the two terms are illustrated in Table 1. Differences Between Surveys and Screening procedures
Principles of Screening Some essential criteria that should be met before instituting a screening programme were postulated by Wilson and Jungner in 1968. They laid out what were later referred to as the 'Ten Commandments of Screening' :
Any screening test contemplated should satisfy the following criteria:
This refers to the proportion of truly diseased individuals in the population who have been correctly identified as diseased by the screening test. It is also refelTed to as the Detection Rate or the True Positivity Rate. A test with a high sensitivity gives only a few false negatives, but many false positives. Specificity: This relates to the proportion of the nonnal individuals who are correctly labeled as non-diseased by the screening test. It is also called the True Nonnal Rate or the True Negative Rate. A test with a high specificity will only give a few false positives, but many false negatives. It is desirable that a screening test should have a high sensitivity and specificity. Though this is theoretically possible, it is practically almost impossible to achieve. If a high sensitivity is desired so that all diseased cases can be identified, the screening test invariably picks up a number of individuals who are non-diseased. This thus increases the number of false positives, which thus means a reduced specificity. Reducing the strictness of the diagnostic criteria increases sensitivity, but lowers the specificity. Glaucoma can bc used to illustrate this point more vividly. Taking the example of vertical cup: disc (C:D) ratios for diagnosing open angle glaucoma, if an arbitrary C:D ratio 'of 0.4 was taken as the cut off between the normal and the glaucomatous cases, most cases of glaucoma would be picked up (high sen.sitivity) but many of the 'so called cases' will turn out to be non-glaucomatous (low specificity). If the C:D ratio of 0.8 is taken as the cut off point, then most of the individuals picked up will definitely be suffering from glaucoma (high specificity), but many of the individuals with glaucoma will actually be missed by this procedure (low sensitivity). The process can be visualized in the following figure.
Predictive value: Predictive values are also very important in validating screening tests. Positive Predictive Value: This indicates the probability of an individual suffering from the disease, when the screening test indicates a positive result. It is desirable to have a high positive predictive value. Negative Predictive Value: This indicates the possibility of an individual not suffering from the disease, when the screening test indicates a negative result. It is desirable to have as high a negative predictive value for a screening test as possible. Following figures illustrate this point.   To decide on whether high sensitivity or high specificity is desirable, infonnation on the consequences of a missed diagnosis is essential. For a condition like refractive errors, high specificity is desirable because cases mis-classified as nonnal would not cause undue hann to the individual (though the credibility of the optometrist will be eroded!). But for a potentially blinding condition like glaucoma it is preferable to have a screening test with high sensitivity, and to bear the expense of a high number of false positives, who will have to be additionally followed up before being labeled as normal. Improving Validity Sometimes, it is seen that a single screening test is insufficient in making even an adequate preliminary diagnosis. Primary open angle glaucoma (POAG) is a shining example of such a condition. For the identification of POAG, no single test is adequate as an effective screening test. The use of either tonometry or ophthalmoscopy alone would result in a high percentage of false negatives. The earlier recommendation was that, as screening with tonometry was a simple, painless and inexpensive procedure, and it took only 5 minutes for each individual, and so could be performed periodically on all asymptomatic adults. It is now known, that nearly half of the POAG patients cannot be detected by tonometry, as they have low intraocular pressures. There is thus a 50% chance that POAG may be missed by a single tonometric measurement, and a further 10% chance that, the patient will be found to have ocular hypertension requiring additional tests, but would not actually have the disease. In these situations, validity can be increased by using the results of several screening tests together. This can be done in two ways: Tests in Series: Here an initial screening test is administered to a population and all those persons who have a positive result, are re-evaluated with an additional test (s). In general, serial testing results in an increase in specificity as compared to a single test, because a positive series is more likely to represent true disease. This system can be used for glaucoma, if perimetry is used as the initial test. and the aim is to prevent the progression of glaucoma to blindness.  Lists in parallel: In a parallel screening system, all the different tests are administered at the same time, and patients with a positive result on anyone test are labeled as abnormal. In general, parallel testing results in increased sensitivity compared with that of each test, since disease is less likely to be missed. However, the procedure obviously has a lower specificity because false positive diagnoses are more likely. Perimetry, ophthalmoscopy and tonometry can be used together for glaucoma, as an example for such tests. but the problem here is the lack of standardized uniform case definitions of the disease, because it not one disease entity, and the relative lack of diagnostic guidelines for all the three approaches. Of course, the costs of such a programme will soar tremendously due to the high number of so called false-positive suspects.  Diagnostic and Treatment Criteria A screening programme as has been emphasized repeatedly, is never to be undertaken as a mere academic exercise. Since it is a service delivery activity with a specific objective of improving the prognosis, it is also important to gauge the treatment and infrastructural facilities available to deal with the apparently abnormal individuals screened out by the programme. Since a screening test is basically a preliminary test that identifies the 'suspects', adequate back up infra-structural facilities should be available to establish a definitive diagnosis. The costs involved in undertaking screening programmes need to be carefully considered. Monetary support is often limited, and it is important to use the scarce resources as optimally as possible. Costing exercises must account for the entire diagnostic work-up that may be necessary to detect a case. According to criteria set out by the Public Health Services Department in USA, any medical screening test must identify a minimum of 2% pathology in the general population to be considered a cost-effective procedure. When the prevalence of a specified condition is low in the general population, yield (number of cases detected) of the screening programme can be increased by targeting specific high-risk groups with a higher prevalence of the disease. Yield is defined as the quantum of previously unrecognized disease that is diagnosed and put on the appropriate treatment, as a result of the screening programme. The positive predictive value is the commonest measure of yield. In this fashion, glaucoma screening is probably more cost effective when targeted at certain subgroups of the population. Here, screening could be offered to the population above the age of 40 years, those with a family history of glaucoma, etc. In comparing costs in screening programmes, the cost per case treated or detected are useful parameters. This also helps in evaluation of the programme. Treatment modalities and protocols on who should be treated and how, should be clearly defined. Once a case is detected, it is the bounden duty of the screening programme to offer the appropriate intervention to the individual. Evaluation of Screening Programmes The value of a screening programme is ultimately determined by its effect on morbidity, mortality and disability. It should never be assumed that a screening programme would definitely benefit the population in all situations. Failure to achieve beneficial results may be related to many inter-related factors. The screening programme can be evaluated in terms of:
Outcome Measures: These measures relate to the effectiveness of the screening programme in reducing mortality, morbidity and disability. It is assumed that early detection leads to a favourable response and improved prognosis, but results should not be accepted at face value. This is because it is well known that people who express their willingness to participate in a screening programme differ to a significant extent from those who refuse to participate. Volunteers for screening programmes are generally more health conscious and tend to get detected early, even without the screening programme, and thus the screening test may not be solely responsible for the detection. Evidence on the role of the screening programme Can therefore be obtained only by undertaking observational or experimental epidemiologic study designs. The best evidence for the effectiveness of screening programmes comes from randomized controlled trials that are a forn1 of experimental epidemiologic designs. Examples of Screening Programmes in Ophthalmic Practice The applications of screening principles are being illustrated with some specific examples in ophthalmic practice. The intention here is to highlight that not all eye problems are amenable to a screening exercise in developing countries at the present moment. Preschool Vision Screening for Amblyopia and Strabismus Studies in the West indicate that the prevalence of amblyopia and strabismus is at least 5%, while only limited population-based data is available from India. Indian data show that it is less of a problem compared to the West. However, there is an attendant disability including unilateral blindness, if the disease is not detected and treated early. Current evidence indicates that normal or near normal binocular vision can be achieved in many cases, if the treatment is completed prior to 24 months of age. Screening for these conditions at the age of 3-5 years or later is not successful. Unfortunately, most of the earlier screening programmes in the West also were initiated at an older age. Effective treatment modalities are available at an early age but as yet there are no agreed policy on guidelines for treatment. One advantage of early screening is that, the consequent blinding complications can be reversed by intervening early in the asymptomatic phase (in terms of blindness). There are many reasons why such a screening programme in developing countries does not seem an attractive proposition, viz.,
The issue of under-referrals or over-referrals assumes a lot of importance for preschool amblyopia screening. Under-referrals of a significant number compromises the effectiveness and credibility of the system, while over-referrals raises false alarms and upsets parents in addition to a loss of credibility. The infrastructure and logistics of treating the screened children is a gigantic task, as very few ophthalmic surgeons have the specialized surgical skills. Even regular refraction and spectacle services are not easily available in the rural areas. In the light of the above, it may not be possible to implement a screening programme at the community level or at the primary level. What is currently feasible, is a system of case-finding or opportunistic screening in the clinics of ophthalmologists at the secondary and tertiary levels, when children present to them with some other problems, or when they accompany their parents or grandparents to the ophthalmologist's office. Glaucoma Screening Programme Glaucoma, the 'lurking thief' has evinced maximum interest from ophthalmologists as a condition for which a screening programme can be instituted. There are valid reasons why ophthalmologists should hold this view. In many countries, it is the second or third important cause of blindness. There are an estimated 22.5 million individuals suffering from Glaucoma in the world of whom 5.3 million are blind. An additional 105 million are suspects having an IOP>21 mm Hg. It is likely to soon become the most important cause of irreversible blindness. The public health significance can not be disputed, and it is also widely recognized that the available figures could be an underestimation. It is generally estimated that the prevalence of Primary Open Angle Glaucoma varies between 1-2 % in the population above the age of 40 years. It is also possible to increase the yield of a screening programme by limiting the screening programme to the population segment above the age of 40 years, and high-risk groups. These risk group include individuals with a family history of glaucoma, those suffering from myopia, central venous occlusion, diabetics, hypertensives, and individuals with thyroid diseases, etc. There is a very high of professional as well as community concern regarding glaucoma and thus would also benefit the programme. However, in spite of all these advantages, glaucoma screening programmes especially in the developing country scenario, are not an easy task. There are many problems related to diagnosis, treatment as well as the infra-structural needs that needs to be addressed for such a programme. These include: Diagnostic criteria: Unlike most of the other disease states, glaucoma is not one disease entity, but a composite mix of different pathologies - Primary Open Angle Glaucoma (POAG), Angle Closure Glaucoma (ACG), Secondary Glaucoma, as well as Congential Claucoma are not one common group of diseases. Thus, establishing uniform case definitions is not possible. It therefore becomes necessary to decide which condition needs to be screened. As against earlier opinion, it is now known, that in most countries, ACG is also important. Recent evidence suggests that even in India, the proportion of POAG to ACG may be a ratio of 1:1. The next problem relates to the actual diagnostic criteria for the diagnostic of any of the glaucoma. Till a few years ago, intra-ocular pressure was accepted as sufficient evidence for the diagnosis of glaucoma. Today, although elevated intra-ocular pressure is clearly the leading risk factor for the development of glaucomatous optic atrophy; the overlap of IOP distribitions in the population with and without glaucoma does not allow for a simple cut off IOP between the normal and apparently abnormal groups. Tonometric screening has been the recommended practice for many years in may countries. But most authorities now feel that, it is questionable whether tonometric screening remains a prudent use of time and resources. If single tonometric readings are obtained in any general population, approximately 10% of the people will have and intra-ocular pressure above the statistical form of 21 mm Hg. However, in actual effect, only a proportion of these individuals will have glaucomatous field loss with an estimated specificity of 10-30%. Additionally, an almost equal number with an intra-ocular pressure (IOP) below 21 mm Hg will have glaucomatous damage, thus giving an overall sensitivity of 50-70%. This means, that exclusive reliance on tonometry would mean that for every patient diagnosed to have glaucoma, approximately 30-50 additional persons who do not have glaucoma must be subjected to further diagnostic tests and unnecessary anxiety. What however is of great concern, is the fact that using IOP as the only criteria for the diagnosis of glaucoma means that nearly 50% of the glaucoma patients may be told that they are free of the disease on the basis of a single tonometric measurement. Some researchers argue that the glaucoma that matters, is the one that is associated with a high intra-ocular pressure that is associated with a rapid progression. and is amenable to control of the high pressure. At the same time, direct ophthalmoscopy alone has also not been found to be very useful as a screening method. There is a Iikelihood of an inter-observer variation, because of the subjective nature of the procedure. Studies have however shown. that direct ophthalmoscopy had a higher detection rate compared to tonometry, but neither method alone was as effective as the two methods combined together. Ophthalmoscopy alone can also gi\'c rise to a high percentage of false positive and negative results. Some studies have also found that the assessment of cup-disc ratios where anything above 0.5 was taken as abnormal, were also insufficient as a screening tool. Mass visual field testing has been shown to be a potentially accurate and efficient means for screening by some authorities. Some studies earlier had found that increased lOP was actually poorly correlated with the presence of glaucomatous field loss. Till recently. perimetry was a very costly option, and was not thought to be suitable for application in the field. Now. newer testing procedures are being tried, including a less expensive manual. hand-held perimeter. However for a country like India. these are not yet a reality. A major problem with field defects is that it is now known that many fibres may already be lost by the time that a functional field defect is detected. This means that the disease would not be actually detected in the early pre-symptomatic phase at which the damage may be reversible. Unfortunately, fields too are associated with numerous false positives. However, the only silver lining is that the number of false positives decrease with learning experience. All these aspects point towards a pessimistic view of instituting a glaucoma screening programme. Many people think that with the current available body of knowledge, if one has to screen for glaucoma, one should use the three tests for the process. Not only is this a costly venture, it has resource implications, in that the available personnel to undertake such testing are limited, and it may not be possible to mount a screening programme in such a situation. It may be prudent to undertake opportunistic screening or case finding at eye clinics in the government and the private sector. At the same time, as the minimum, any screening programme in captive groups like factory workers, need to employ a minimum of two tests. The two initial tests which can thus be utilized are lOP (>2lmm Hg) and the vertical cup disc ratio (>0.5). If an individual tests abnormal on either of the two tests, then the visual fields should be undertaken. If the attempt is to screen angle closure glaucoma, or to differentiate between the two primary adult types, gonioscopy facilities should be available at the secondary level at least. If a hand-held, inexpensive and easily administered method for examination of the fields is available, it may ultimately become an integral part of the screening programmes. Treatment modalities: In addition to the diagnostic jigsaw puzzle. is the question of treatment. Clear cut guidelines on treatment policies will have to be established. Compliance and the costs of tr~atment also need to be considered for this purpose. At the present juncture, it appears that medical treatment may not be a teasible alternative for most populations in the developing countries, especially those residing in the rural areas. Long term compliance as well as the therapeutic effects of medical treatment are now being questioned. The world slowly seems to veer around to the view that primary surgery may be the best alternative. Recently, two long-term studies compared the efficacy of different treatment modalities including medical treatment, argon laser trabeculoplasty and primary trabeculectomy. These studies showed that the primary trabeculectomy group had the best results and this procedure was associated with the lowest lOP postsurgically and the long-term consequences in terms of field loss were minimal with the surgical option. This points to the fact that in most developing countries, it may be best to provide primary trabeculectomy as the treatment of choice. This may ultimately be achieved at the peripheral surgical facilities. Diabetic Retinopathy Screening Programme Till very recently, diabetes mellitus was thought to be a disease which afflicted the affluent populations in the developed world. Recent evidence shows that diabetes is now a major problem in many developing countries also, especially in the rapidly expanding urban areas of these countries. One most interesting important observation is that, if an arbitrary cut off of 10% is accepted as indicative of a high prevalence of diabetes, then this is seen only in the populations in the developing countries or migrants from these ethinic backgrounds to the developed countries. In India, the prevalence is estimated to be approximately 1 % in the general population in the rural areas, and nearly 11 % in the urban area. It is also known that diabetics are 10-20 times more likely to go blind, compared to the non-diabetics. 80% of diabetics have a retinopathy after 20 years, although the retinal changes become evident at an earlier stage in the maturity onset, noninsulin dependant diabetics. In the United Kingdom, 50% of the diabetic blind are dead within 3-4 years of registration, and only about 20% survive for about 10 years. It is thus prudent to think of a programme for screening in diabetics, for preventing blindness from retinopathy. Screening strategies depend on the late of appearance and progression of diabetic retinopathy and on the risk factors that alter these rates. In general, the progression of retinopathy is an orderly process extending from mild background abnormalities, advancing to pre-proliferative retinopathy and ultimately to proliferative retinopathy. The prime motivation for screening for diabetic retinopathy is the established efficacy of laser photo-coagulation for treatment. However, the primary screening needs to be undertaken using the non-ophthalmologists, as diabetes is a multi-system disease that needs dietary and diabetic control with medication, in addition to the other systemic complications. Though in the West, stereoscopic fundus photographs are the gold standard for diagnosis, the costs involved do not allow the same to be extrapolated to the developing countries. If general duty medical officers or optometrists are to be involved, ophthalmoscopy, on a dilated pupil may be the only method, but this is a highly insensitive procedure. It may be useful to screen populations above the age of30 years for diabetes_, especially in the presence of a family history, and to refer such patients to an ophthalmologist if there are any abnormal fundus features on dilated ophthalmoscopy. At the present time, it does not appear that screening for diabetic retinopathy is possible at the primary level for the general population,-but all general physicians should be involved in an opportunistic case finding procedure at the clinic level or in captive population groups. However recent studies in the country have indicated that rigorous training of paramedical staff can prove usefull in screening. Certain guidelines have been set out by the American College of Physicians, American Diabetes Association and the American Academy of Ophthalmology. These include: Patients with type I diabetes should be screened annually for retinopathy, beginning 5 years after the onset of diabetes (No need for screening before puberty). Patients with type II diabetes should have an initial examination for retinopathy shortly after the diagnosis of diabetes is made. If dilated ophthalmoscopy is used, then the examinations should be repeated annually. Women with diabetes mellitus who become pregnant, should have a comprehensive eye examination in the first trimester, and a close follow up throughout pregnancy. Gestational diabetics, however do not need such a screening, because there is no increased risk in such women. 'Screening' for refractive errors in school children Refractive errors are a major cause of ocular morbidity in India. The prevalence of refractive errors in the general population has been estimated to be 14.2%.7.35% of bilateral blindness, 60% of one-eyed blindness, and 18.87% of low vision are due to refractive errors in India. In terms of proportional morbidity, 46.69% of all ocular morbidity in the country is directly attributable to refractive errors. As far as the younger age group is concerned, it is known that myopia is the commonest of the refractive errors. It is generally believed that children start developing myopia when they enter school, and progression usually stops when the child grows up to the school leaving stage. Children who acquire myopia early or are born with high myopia are handicapped to a great extent, and the progression of myopia is much faster among them. In addition to the high prevalence of refractive errors, is the fact that they also lead to a loss of productivity and efficiency, leading to a reduced work output.It is known that poor vision in childhood could adversely affect academic performance which could then lead to a lack of interest in studies, and consequently to school drop-outs. The justification for a programme targeted for the school children comes from the fact that this captive group of children is an important human resource for the future, and investing in their ocular health pays rich dividends over time. The major reasons why such programmes have been undertaken are
For the initial vision screening, a single optotype of the Snellen's chart or the 'E' chart can be easily administered by minimally trained personnel like school teachers. This is a low cost, non-invasive, rapid, reliable and acceptable method. The conventional Snellen's Charts with all the 7 lines of optotypes may be confusing for administration by personnel like the school teachers, and should not be used. In addition, the conventional charts are easily memorized by the children, and this would not be useful for screening. A Single optotype like the 'E' can be rotated each time the child sees it, and thus each eye can be tested differently. DANIDA, in India, has developed a tool including 4 'E' optotypes of 6/9 line with the limbs of the 'E' facing in different directions. Children are then asked to identify at least three optotypes with each eye (rotating the card for the second eye, so that the letters are in a different configuration) before labeling them as abnormal or normal. There are some pertinent issues to be considered before embarking on a vision screening programme among school children. These include: A good back up system for diagnosis and refractive correction is necessary. This system would include ophthalmic Assistants/Technicians at the periphery supported by the Ophthalmic Surgeons for tackling the difficult or compound refractive errors. The logistics for the provision of spectacles should be carefully considered. This includes the availability of opticians as well as the costs of providing the spectacles. In India. it is estimated that the cost of providing a pair of spectacles to a child after a screening programme is Rs. 250 per corrected child. The entire screening exercise works out to Rs. 5.00 per each child screened under the programme. The frequency of screening also needs to be carefully deliberated. In general, most authorities opine that screening once at school entry (or at the primary school stage) and once again at the entry to the middle/ secondary school is required. Thus, each child should be screened at least twice at a minimum interval of 4-5 years. There is a lot of debate on whether instituting a programme for 10-12 years old children constitute a screening programme. and the utility of such an approach. Some people feel that at this late age, most children would be aware of their handicap, and may themselves approach a health facility for redressal of their problem. Thus they feel that this has a low utility in terms of the costs involved, and should not be labeled as a screening process, but merely as a case finding approach. It then means that screening needs to be undertaken much earlier at the school entry stage. However, this is fraught with two major problems. The first relates to the vision screening methodology to be adopted at this age, and the second relates to the difficulty in terms of the logistics involved in initiating such a service as the number of school children at this stage is enonnous. The effect of the early detection and correction of refractive errors in terms of the final refractive power attained by the individuals is not yet clear. The question basically relates to the natural history and progression of uncorrected refractive errors, and as to what happens to the final myopia if the child remains uncorrected. Another problem relates to the compliance with spectacles. Generally it is stated that compliance with regular use of spectacles is poor in most communities in the developing countries due to the stigma attached to use to spectacles. This is particularly so in relation to children, especially the girl child. Thus correcting children with a marginal handicap may not work, as such children will have less motivation to wear their spectacles constantly. Children with a visual acuity of <6/12 in the better eye may be better candidates to comply rather those with better levels of vision. The most important aspect of such a screening programme relates to the validity of the screening test. This is closely related to the cut-off point of visual acuity selected. If a visual acuity of <6/9 in the better eye is targeted, the method will be highly sensitive, but will have a low specificity and will result in a large number of false positives. Ifhowever, on the other hand, a visual acuity of< 6/18 in the better eye is selected as the cut off point for abnormal vision, the test would be extremely specific as almost everybody picked up as abnormal would be really suffering from a significant refractive error. But a few cases of refractive error will tend to be missed, and thus, the screening test will have a lower sensitivity. A compromise trade-off may be an acuity of <6/12 in the better eye. It is necessary to experiment with different cut off points to select the optimal trade-off in different communities. This obviously has to be related to the level of community concern, i.e. at what level of visual handicap is a community willing to accept correction services. A programme like the school vision screening programme needs to be evaluated in terms of the impact of the early intervention. Evaluation of the benefits of a vision screening programme in school children should address the following issues:
The above examples show that there is no single eye disease that can satisfy all the criteria for screening. The decision on which disease needs to be targeted needs to be carefully considered using technical, logistic and cost information. School Vision Screening Programme Component Under NPCB Vision Screening of school children is now one of the 4 major component activities under the 100% centrally sponsored National Programme for Control of Blindness (NPCB). The approach was first tried out in the 5 pilot districts adopted by DANIDA, during the second phase of its assistance to the NPCB. It has now been extended to the whole country. To strengthen the school health service component, the government has proposed that an ophthalmic assistant should be posted in all Primary Health Centres (PHC) covering a population of30,000. Ultimately, it has been envisaged, that dark room facilities will be created at each of these PHCs. Currently a PMOA is only posted at Community Health Centre (CHC) (1: 100000 population). The programme will be undertaken by the DBCS. Funds will be earmarked for this component in the annual plan of the district. Since most of the districts in the country are big, it is prudent to cover the most backward and remote areas of the district in the first instance, and to slowly cover the rest of the district. The process may take up to 5 years depending on the size, resources and the availability of PMOAs in the specific district. Under the NPCB, the following are the important issues related to the school screening programme.
|
|
|
