Methods

The RAND-like appropriateness methodology applied in this study was similar to that used in a similar study in the United States and used all 148 polling statements created for the United States study. Ten additional statements pertaining to fixed-combination glaucoma therapy available in Latin America, but not in the United States, were added, for a total of 158 polling statements. The 10 novel statements were created by one of the authors (S.F.L.) and were reviewed by 2 additional steering committee members with previous experience using the RAND-like method (K.S. and M.R.W.). A fourth steering committee member (R.S.) was not involved in creating the polling statements and thus was allowed to serve as a voting panelist. The voting panel consisted of 8 glaucoma specialists. Members of the panel were selected primarily because of recognition of them as leaders in the Latin American glaucoma community, with fluency in English, the language used for all study communication, being an important prerequisite qualification. There were 2 voting panel members each from Argentina, Brazil, and Colombia, as well as one each from Chile and Mexico. All panelists were required to attend the face-to-face panel meeting for methodologic reasons.

The polling statements were divided into the following categories: medical therapy, adjunctive medical therapy, and assessment and modification of medical therapy; laser trabeculoplasty; glaucoma surgery; trabeculectomy; combined cataract and trabeculectomy; aqueous drainage devices; nonpenetrating glaucoma drainage surgery; diagnostic testing; and general considerations. The panelists were provided with a comprehensive list of articles pertaining to the polling statements, with an effort made to select literature with the highest possible level of evidence based on criteria from the Oxford Center for Evidence-Based Medicine Levels of Evidence and Grades of Recommendation. The list included only articles that met Oxford grading level 1 or 2 and that had been published within the preceding 15 years.

In addition to the mandatory review of publications meeting the above criteria, the panelists were free to review any other published material, as they deemed appropriate, regardless of whether the work was carried out in Latin America. A series of syntheses dealing with major topics addressed in the polling statements, previously developed by the methodologist for the United States study (B.L.L.) were made available to the voting panelists. The polling process began with a list of statements and all relevant articles and syntheses being electronically sent to each panelist as an Excel file (Microsoft Corporation, Redmond, Washington, USA). Instructions on grading polling statements were provided to the panelists. The polling scale was identical to that used by the United States Rand-like Appropriateness Methodology Group. Briefly, a scale of 1 through 9 was applied to assess the validity of a statement: grades of 1 through 3 were considered inappropriate or false, grades of 4 through 6 were considered indeterminate or equivocal; and grades 7 through 9 were considered appropriate or valid. Ratings of 1 and 9, respectively, signified strongest disagreement and strongest agreement with the polling statements. The panelists were given 2 weeks to review all documents, after which they submitted responses electronically to a statistician who analyzed the results. The statistician was not a member of the steering committee or the voting panel, and further had no financial relationship with any of the sponsors of the study or manufacturers of any products referenced directly or indirectly in the polling statements. None of the sponsors had access to, or influence over, the polling statements, selection of panel members, meeting location or date of the panel, or results of this study. Sponsors, or their representatives, were not allowed to attend any meetings related to this project. The steering committee was masked with regard to the individual responses of the panelists at all times during the course of the process. Five weeks after the initial polling, all participants were invited to a panel meeting to discuss the results and to clarify any misinterpretations regarding the wording or interpretation of individual polling statements. Although the results of the similarly conducted United States RAND-like glaucoma project had been published before this meeting, there was no reference to this earlier work during the Latin American panel meeting. The collective results pertaining to all 158 polling statements were reviewed by the voting panel members in attendance without knowledge of how any individual panelist voted for each statement. All panelists attending the panel meeting voted again within 2 weeks of the meeting, using the same methodology as in the initial voting process. The results of this second polling were considered final and are presented here.

Results

All 8 panelists completed the premeeting statements as requested, attended the meeting, and also submitted their postmeeting responses. There was consensus agreement or disagreement for 51.3% (81 of 158) and 66.5% (105 of 158) of the polling statements in the prepanel and postpanel meeting voting, respectively. Eighty-three (79.0%) of 105 of the postmeeting consensus statements where consensus was reached represented agreement with the polling statement, with the remaining 22 representing disagreement. In the United States study, the panel reached consensus in 55.4% of statements, increasing to 74.3% of statements after the meeting (71.8% consensus agreement).

The Supplemental Table (available at AJO.com ) summarizes the Latin American and United States results for the 148 polling statements used in both studies, with the abbreviations A, D, and I representing consensus agreement, consensus disagreement, and indeterminate responses, respectively.

There were some notable differences between glaucoma practitioners from the United States and those practicing in Latin America. In the medical therapy section, the United States panel agreed to consider topical carbonic anhydrase inhibitors (CAIs) or α-agonists as acceptable for first-line therapy, whereas the Latin American panel was indeterminate. The United States experts disagreed and the Latin American experts were indeterminate regarding the use of a fixed combination agent (topical CAI and β-blocker) as first-line therapy. The United States panel disagreed and the Latin American panel was indeterminate with regard to the statement that systemic safety is greatest with CAIs and in considering that ocular side effects are least severe with topical CAIs. Although the United States panel reached consensus agreement that medical therapy should be initiated with a 1-eye trial, the Latin American panel was indeterminate on this issue. The Latin American panel deemed the availability of a drug sample in the office as being less important in choosing a therapeutic agent than the United States panel and was indeterminate regarding consideration of race or ethnicity in such a choice, whereas the United States panel disagreed with this latter polling statement. The United States panel agreed and the Latin American panel was indeterminate with regard to the importance of age as a factor for choosing a therapeutic class of drugs.

No differences were found between panels regarding indications for laser trabeculoplasty and glaucoma surgery. Whereas the Latin American panel agreed that mitomycin C should be used routinely as an adjunct with combined cataract and glaucoma procedures, the United States experts were indeterminate on this issue. Differences in the use of glaucoma drainage devices (GDDs) included the Latin American panel agreeing to consider GDD implantation as an alternative to trabeculectomy, with the United States panel being indeterminate on this issue; further, the United States panel disagreed that the complications with GDD implantation are similar to those found with trabeculectomy, with the Latin American panel being indeterminate on this issue. The Latin American and United States panels were indeterminate and disagreed, respectively, regarding all GDDs having similar success rates. The United States panel disagreed in considering nonpenetrating glaucoma surgery as an acceptable alternative surgery to trabeculectomy for maximal lowering of IOP and agreed in that nonpenetrating glaucoma surgery has a lower postoperative complication rate than trabeculectomy, with the Latin American panel being indeterminate with regard to both of these statements.

Regarding diagnostic testing, the Latin American panel agreed that stereo disc photographs are more useful than nonphotographic imaging of the disc for diagnosing glaucoma, whereas the United States panel was indeterminate on this issue. The United States panel disagreed and the Latin American panel was indeterminate in considering standard automated perimetry (Swedish interactive threshold algorithm) as the gold standard for diagnosing glaucoma. The United States experts agreed that frequency doubling perimetry can detect glaucoma or glaucoma progression earlier than standard automatic visual field testing and that central corneal thickness should be measured in all POAG patients, whereas the Latin American panel was indeterminate with regard to these 2 statements.

The Latin American panel agreed in advancing therapy whenever optic nerve or visual field progression is confirmed, whereas the United States panel was indeterminate with regard to these statements. There was 1 polling statement for which there was absolute discordance between the Latin American and the United States panels. The Latin American and United States panelists reached consensus agreement and disagreement, respectively, with the necessity of advancing therapy when IOP consistently exceeded the target pressure, even without documented optic nerve or visual field progression. The Latin American panel agreed that it is useful to measure systemic blood pressure and same-day IOP curves in POAG patients and that patient difficulties with activities of daily life should be assessed as part of glaucoma assessment. The United States panel was indeterminate with regard to these 3 statements.

The United States and Latin American panels reached consensus disagreement and were indeterminate, respectively, regarding 18 mm Hg being a target pressure for a POAG patient with an early nasal step (−2 dB).

Results

All 8 panelists completed the premeeting statements as requested, attended the meeting, and also submitted their postmeeting responses. There was consensus agreement or disagreement for 51.3% (81 of 158) and 66.5% (105 of 158) of the polling statements in the prepanel and postpanel meeting voting, respectively. Eighty-three (79.0%) of 105 of the postmeeting consensus statements where consensus was reached represented agreement with the polling statement, with the remaining 22 representing disagreement. In the United States study, the panel reached consensus in 55.4% of statements, increasing to 74.3% of statements after the meeting (71.8% consensus agreement).

The Supplemental Table (available at AJO.com ) summarizes the Latin American and United States results for the 148 polling statements used in both studies, with the abbreviations A, D, and I representing consensus agreement, consensus disagreement, and indeterminate responses, respectively.

There were some notable differences between glaucoma practitioners from the United States and those practicing in Latin America. In the medical therapy section, the United States panel agreed to consider topical carbonic anhydrase inhibitors (CAIs) or α-agonists as acceptable for first-line therapy, whereas the Latin American panel was indeterminate. The United States experts disagreed and the Latin American experts were indeterminate regarding the use of a fixed combination agent (topical CAI and β-blocker) as first-line therapy. The United States panel disagreed and the Latin American panel was indeterminate with regard to the statement that systemic safety is greatest with CAIs and in considering that ocular side effects are least severe with topical CAIs. Although the United States panel reached consensus agreement that medical therapy should be initiated with a 1-eye trial, the Latin American panel was indeterminate on this issue. The Latin American panel deemed the availability of a drug sample in the office as being less important in choosing a therapeutic agent than the United States panel and was indeterminate regarding consideration of race or ethnicity in such a choice, whereas the United States panel disagreed with this latter polling statement. The United States panel agreed and the Latin American panel was indeterminate with regard to the importance of age as a factor for choosing a therapeutic class of drugs.

No differences were found between panels regarding indications for laser trabeculoplasty and glaucoma surgery. Whereas the Latin American panel agreed that mitomycin C should be used routinely as an adjunct with combined cataract and glaucoma procedures, the United States experts were indeterminate on this issue. Differences in the use of glaucoma drainage devices (GDDs) included the Latin American panel agreeing to consider GDD implantation as an alternative to trabeculectomy, with the United States panel being indeterminate on this issue; further, the United States panel disagreed that the complications with GDD implantation are similar to those found with trabeculectomy, with the Latin American panel being indeterminate on this issue. The Latin American and United States panels were indeterminate and disagreed, respectively, regarding all GDDs having similar success rates. The United States panel disagreed in considering nonpenetrating glaucoma surgery as an acceptable alternative surgery to trabeculectomy for maximal lowering of IOP and agreed in that nonpenetrating glaucoma surgery has a lower postoperative complication rate than trabeculectomy, with the Latin American panel being indeterminate with regard to both of these statements.

Regarding diagnostic testing, the Latin American panel agreed that stereo disc photographs are more useful than nonphotographic imaging of the disc for diagnosing glaucoma, whereas the United States panel was indeterminate on this issue. The United States panel disagreed and the Latin American panel was indeterminate in considering standard automated perimetry (Swedish interactive threshold algorithm) as the gold standard for diagnosing glaucoma. The United States experts agreed that frequency doubling perimetry can detect glaucoma or glaucoma progression earlier than standard automatic visual field testing and that central corneal thickness should be measured in all POAG patients, whereas the Latin American panel was indeterminate with regard to these 2 statements.

The Latin American panel agreed in advancing therapy whenever optic nerve or visual field progression is confirmed, whereas the United States panel was indeterminate with regard to these statements. There was 1 polling statement for which there was absolute discordance between the Latin American and the United States panels. The Latin American and United States panelists reached consensus agreement and disagreement, respectively, with the necessity of advancing therapy when IOP consistently exceeded the target pressure, even without documented optic nerve or visual field progression. The Latin American panel agreed that it is useful to measure systemic blood pressure and same-day IOP curves in POAG patients and that patient difficulties with activities of daily life should be assessed as part of glaucoma assessment. The United States panel was indeterminate with regard to these 3 statements.

The United States and Latin American panels reached consensus disagreement and were indeterminate, respectively, regarding 18 mm Hg being a target pressure for a POAG patient with an early nasal step (−2 dB).

Discussion

Although there have been many large multicenter clinical trials to support decisions in glaucoma practice in recent years, many and perhaps most decision making remains based on clinical impressions and expert opinion. The information from trials does not address all aspects of glaucoma management, and even when such evidence is available, one must consider that the subject profiles of those randomized in such trials are not always similar to patients seen in routine practice. Consensus methodology has been developed in an effort to merge available evidence-based data with expert opinion. The modified RAND-like approach is one such methodology that is believed to have a significant impact on practice patterns. This modification of the RAND method has been applied previously to the management of glaucoma and glaucoma suspects in various regions of the world.

The population of Latin America and the Caribbean is projected to increase from 580 million in 2009 to 724 million in 2050. This projected increase partially reflects the improvement in health care during the latter half of the twentieth century, which has led to declining death rates in developing and developed countries in the region. The population and life-expectancy increase, combined with the importance of age as a risk factor for glaucomatous disease, undoubtedly will lead to glaucoma becoming an increasingly important public health problem in Latin America over the coming decades. Using models based on prevalence studies performed on Hispanic residents of the United States, it has been estimated that by the year 2020, 12.9% of the global open-angle glaucoma population will reside in Latin America. Currently, there is no specific information available in the peer-reviewed literature pertaining to glaucoma diagnosis and treatment in this region of the world. The modified RAND- like methodology presented in this article reflects the current glaucoma practice patterns for open-angle glaucoma in Latin America.

Some notable highlights of the panel results include agreement with the use of prostaglandin analogs as the preferred first-line agents for glaucoma therapy, with other classes deemed acceptable for initial therapy. Compliance with therapy was determined to be inversely related to the number of eyedrops used each day. Difficulties in the assessment of compliance with medications for chronic conditions such as glaucoma persist and, although less frequent daily administration of drops is considered to be desirable, other specific medication-related factors also may play an important role in influencing compliance.

The panel was indeterminate with regard to the usefulness of a 1-eye trial when beginning treatment, adding an agent, or reducing the number of medications. Although such a trial is supported by the Practice Preferred Patterns of the American Academy of Ophthalmology and also is mentioned in the European Glaucoma Society Guidelines, evidence in support of this approach is conflicting, and concordance in IOP variability among fellow eyes may not be as reliable as once thought. The factors found to be most relevant to choosing a therapeutic class of drugs related to the efficacy and tolerability of available agents. Costs to the individual patient were deemed more important than cost to society in Latin America. The availability of samples was not deemed to be important in choosing therapy. These results suggest that patient well-being is far more important than convenience to the practitioner, or impact on society, when therapeutic options for glaucoma care are weighed in Latin America.

The panel generally agreed on the definitions of noncompliance with therapy and the value of setting a target IOP goal when commencing therapy. There was also agreement that when there is nonresponse to initial therapy, switching classes or switching within a class is preferable to adding another medication. This approach is certainly more cost effective for patients than adding agents to existing therapies that are ineffective. Once again, the Latin American experts voted in favor of what is optimal for individual patients rather than what is most convenient for the practitioner. In terms of advancement of treatment, the panel voted to proceed to the next therapeutic step when disease progression was noted on any of a number of parameters or, when a predetermined target IOP goal had not been reached, even in the absence of progression. These voting results reflect a willingness to err on the side of overtreatment versus undertreatment, an approach more consistent with developed rather than developing countries. This may be a reflection of the fact that the overwhelming majority of the Latin American panelists practice in affluent urban or suburban settings where high-quality care comparable with that seen in the United States and Europe is readily available to those who can afford it. Latin America is composed of nations and regions within these nations that can be considered developing or developed, and thus it must be acknowledged that the results of this panel may present practice patterns that are skewed toward those of developed countries because of glaucoma leaders generally practicing in such an environment.

One topic that has not been well studied previously in Latin America is the use of fixed combinations as first-line agents. In particular, prostaglandin–β-blocker fixed combination agents are not available in the United States and some other developed countries. It is noteworthy that the panel disagreed with the use of the prostaglandin–β-blocker analog as first-line fixed combination therapy, perhaps reflecting a viewpoint that a prostaglandin alone may be almost as efficacious in terms of IOP lowering without the added costs and systemic risks associate with β-blocker therapy.

With regard to laser trabeculoplasty, the Latin America panel did not recommended it as a first-line option for treatment of POAG, but agreed that it should be used as an alternative to second- or third-line therapy. Although the results of the Glaucoma Laser Trial determined that initial treatment with argon laser trabeculoplasty was at least as efficacious as initial medical treatment, this study was conducted before the availability of several contemporary classes of glaucoma medications, and the panel was not convinced that laser trabeculoplasty is presently an ideal first-line therapeutic choice for POAG. Argon and selective laser trabeculoplasty were considered equivalent in terms of efficacy and short-term side effects. Results were indeterminate with reference to their relative effectiveness in the retreatment of the angle. In general, the voting results suggest that the introduction of selective laser trabeculoplasty has not impacted the glaucoma treatment algorithm significantly for care in Latin America.

There was generally strong agreement among the panelists in the glaucoma surgery category. It seems that despite the increasing popularity of nonpenetrating and drainage device procedures globally, trabeculectomy remains the gold standard surgical procedure for glaucoma in Latin America. There was also agreement that antifibrotic agents should be used routinely as adjuncts with trabeculectomy, and in the case of combined glaucoma and cataract procedures, mitomycin C was deemed preferable to 5-fluorouracil. However, consensus could not be reached regarding the relative merits of 1-site versus 2-site combined procedures. GDDs were considered an alternative to trabeculectomy in most cases, although it was agreed that lower IOP can be obtained with mitomycin C–augmented trabeculectomy. It is noteworthy that the 3-year postoperative follow-up report of the Tube versus Trabeculectomy Study showed a higher success rate for the tube group with similar mean IOP and number of postoperative medications, but higher cumulative probability of failure in the trabeculectomy with mitomycin C group relative to the drainage device group. The overall incidence of complications also was higher in the trabeculectomy group in this study. Despite these findings, it seems that Latin American practitioners generally prefer trabeculectomy over drainage device implantation, particularly as an initial surgical procedure for glaucoma refractory to medical and laser therapy.

The group agreed that disc photographs are needed to manage glaucoma and that stereo disc photographs are preferable. The preference for stereo disc photographs over imaging devices was in concurrence with previous studies, although some reports have shown comparable results between these 2 methods to measure structural damage related to glaucomatous disease. The growing global use of optic nerve imaging methods using laser or optical coherence tomography have impacted significantly the choice of test to follow up structural changes in the optic nerve in developed countries. By voting in favor of stereo disc photographs as the gold standard in this arena, the panel results may reflect the fact that these new technologies, although promising, should not replace stereo optic nerve photography at the present time in Latin America. In general, the panel’s views on the frequency of required structural and functional assessment of the optic nerve as well as gonioscopy were similar to those recommended by the American Academy of Ophthalmology in their Preferred Practice Pattern document. There was strong agreement that the presence of structural damage, even without functional loss, may be considered an indication to initiate IOP-lowering treatment. A consensus was reached that progression, either structural or functional, is an indication to advance therapy. The relevance of preperimetric glaucoma has been debated widely among glaucoma practitioners and regulatory agencies throughout the world. The results of this panel suggest that Latin American glaucoma experts recognize the importance of diagnosing glaucoma in the earliest stages, sometimes before visual manifestations on automated perimetry.

In comparing the results of this Latin American RAND-like methodology study with a previously published similar study in the United States, it is noteworthy that there was strong concordance between the 2 studies, with no differences found for 111 (75%) of 148 statements and complete discordance found for only 1 statement relating to advancement of therapy in the absence of documented progression in the instance of IOP being higher than target, with the Latin American group advocating a more aggressive IOP-lowering strategy than their United States colleagues. One major reason for such strong overall concordance may be that most Latin American panelists in this study practice in urban areas with greater wealth and access to care than found in many rural regions of the continent, and thus their practice patterns may be more likely to resemble those of United States practitioners than would be expected if the Latin American panel consisted predominantly of rural practitioners.

Latin America is a large region of the world, the practice patterns of which cannot be represented fully by the handful of experts involved in this project. Every effort was made to assemble a panel including individuals of diverse backgrounds with regard to glaucoma training and region of practice. Although not perfect, the views of these experts pertaining to glaucoma care in Latin America are more relevant to that region than the views of similarly trained individuals practicing in North America and Europe. Region-specific practice patterns are best assessed by engaging those who are familiar with the diagnosis and management of glaucoma in the areas being evaluated. We are hopeful that this study provides a summary of such practice patterns to those involved in glaucoma care in Latin America and that it will serve as a stimulus for other comparative assessments in this region and throughout the world.