- •
All published literature does not have equal import.
- •
All medical literature, including this book, must be read critically.
- •
Whenever possible, treatment decisions should be based on evidence-based medicine. Evidence-based medicine is defined as the conscientious, explicit and judicious use of current best evidence in making decisions about the care of individual patients.
- •
The best evidence usually comes from well-conducted, randomized clinical trials; however, there are relatively few randomized clinical trials in uveitis, although the number is increasing.
- •
It is important to understand the strength or level of evidence supporting treatment decisions.
The practice of medicine is based on applying what we have learned from our medical training in addition to the knowledge gained by reading books and articles and attending scientific conferences. It is our hope, for example, that the book you are reading will assist you in the care of patients with uveitis. However, all medical literature, including our book, must be read critically. Much of the information that appears to be irrefutable scientific dogma is actually based on inconclusive data derived from a handful of patients. Because the recommendations are in print and stated by seemingly reputable authorities, they are often followed blindly. It is frequently useful to thoroughly review the original references on which various therapeutic approaches are based; what you find may surprise you.
There is a growing movement to basing treatment decisions on evidence-based medicine. Evidence-based medicine is defined as the conscientious, explicit and judicious use of current best evidence in making decisions about the care of individual patients. Evidence-based medicine has been divided into two types: evidence-based guidelines and evidence-based individual decision making. In each case, the goal is to base therapy on the best evidence available. A number of classifications for the quality of evidence have been proposed. One commonly used stratification was developed by the US Preventative Services Task Force and is detailed in Table 6-1 ). An assessment of the quality of the information in the literature is critical in determining the best treatment for our patients.
| Level I | Evidence obtained from at least one properly designed randomized controlled trial |
| Level II-1 | Evidence obtained from well-designed controlled trials without randomization |
| Level II-2 | Evidence obtained from well-designed cohort or case-control analytic studies, preferably from more than one center or research group |
| Level II-3 | Evidence obtained from multiple time series with or without the intervention. Dramatic results in uncontrolled trials might also be regarded as this type of evidence |
| Level III | Opinions of respected authorities, based on clinical experience, descriptive studies, or reports of expert committees |
When treatment recommendations are made in published guidelines or in the literature, they can also be categorized by the level of evidence on which the information is based. The US Preventative Services Task Force uses the categories listed in Table 6-2 .
| Level A | Good scientific evidence suggests that the benefits of the clinical service substantially outweighs the potential risks. Clinicians should discuss the service with eligible patients |
| Level B | At least fair scientific evidence suggests that the benefits of the clinical service outweighs the potential risks. Clinicians should discuss the service with eligible patients |
| Level C | At least fair scientific evidence suggests that there are benefits provided by the clinical service, but the balance between benefits and risks are too close for making general recommendations. Clinicians need not offer it unless there are individual considerations |
| Level D | At least fair scientific evidence suggests that the risks of the clinical service outweigh potential benefits. Clinicians should not routinely offer the service to asymptomatic patients |
| Level I | Scientific evidence is lacking, of poor quality, or conflicting, such that the risk versus benefit balance cannot be assessed. Clinicians should help patients understand the uncertainty surrounding the clinical service |
When you read the medical literature, it is important to determine the experimental methods used in the various studies. Sometimes this information can be discerned from the title of the paper; at other times the Methods section must be read. Occasionally the study method is not stated in the paper: if this is the case, the article is probably not worth reading. Many of the clinical studies in the literature are retrospective, meaning that the data were collected from previous patient visits. Retrospective studies can provide valuable information but are usually limited by the quality and thoroughness of the patient records. A patient’s symptoms and clinical findings, although present, are often not recorded. For example, if there is no comment in a patient’s record about vitreous haze, does this mean that vitreous haze was absent or that it was present but not recorded in the note? Fortunately, most ophthalmic notes contain a core of data including visual acuity, intraocular pressure, and results of anterior segment and retinal examination. Nevertheless, prospective studies, when specific data are collected during patient visits on specifically designed case report forms, are less prone to errors.
Study design
There are four basic clinical studies: case series, case–control studies, cohort studies, and randomized clinical trials. The case report or case series is probably the weakest method of deriving clinical data. Case series are usually retrospective reviews that list the clinical findings of patients with a specific disease. Case series can illustrate the variety of clinical manifestations of diseases and provide information about diagnosis, management, and prognosis. However, in addition to the problem of data missing from patient records, the reader should be aware of other pitfalls that may jeopardize the value of the report. First, the disease or condition may not be adequately defined. If patients with sympathetic ophthalmia are inadvertently included in a series of patients with Vogt–Koyanagi–Harada syndrome, the findings and conclusions may be altered. Second, the patient population reported may be dissimilar from that in the clinician’s practice. Frequently uveitis specialists see patients with more severe diseases because these are the ones referred to their practice. Therefore, a uveitis specialist may report that the visual prognosis for a given disease, such as sarcoidosis, is poor. If this report is based on a series of patients composed of referred patients with end-stage disease, it may be biased. Be especially wary of retrospective reviews that make global dogmatic statements on the basis of a few patients. Finally, case series have no control group for comparison. If a report states that depression was found in 35% of patients with uveitis, it is not clear that this represents a causal relationship. Visual loss and not uveitis may be the cause of depression. It would be important to know how many patients with other ocular diseases that cause visual loss (a control group) have depression. For example, the same percentage of patients with visual loss from retinal degenerations may also be depressed.
The case–control study is a second type of clinical study in which the investigator compares a group of patients with a given condition to a control group without the condition. The records of both groups are then compared to see whether certain factors were more likely to occur in one group than in the other. The classic example of a case–control study would be to examine patients with lung cancer and a group without the disease and determine their smoking history. It is then possible to compute an odds ratio that determines the relative risk for a given condition such as lung cancer, given a specific factor such as smoking. Case–control studies, albeit more powerful than case series, are prone to bias, many relying on a retrospective review of patients’ records to determine the differences in a number of clinical parameters between cases and controls. Additional bias arises from the method of choosing the case and control subjects. Despite the potential bias, case–control studies are becoming more common in the literature because they are easy to carry out, especially with the computerization of clinical records. In addition, they are often the only feasible method available to study rare disorders.
A cohort study identifies two groups or ‘cohorts’ of patients, for example one cohort that receives a certain treatment and one cohort that does not. The groups are then followed prospectively for the development of a specific outcome. However, because the treatments are not assigned randomly, it is possible that the two groups differ in important parameters. For example, if you followed two groups of patients with glaucoma, one treated with medications and one treated surgically, you may falsely conclude that surgical therapy is inferior for treating glaucoma because these patients had a worse visual outcome. However, it may be that the patients treated medically had milder disease and better initial visual acuity than patients who received surgical therapy.
Pharmaceutical drug development includes a number of clinical studies, but the final determination of safety and efficacy is based predominantly on pivotal randomized, controlled clinical trials. Clinical studies during the development of a new medication are divided into four phases. Phase I clinical trials are the initial safety trials of a new medicine and are usually conducted in normal volunteers. The trials can be open label, for which patients and investigators are unmasked to the treatment allocation. Multiple doses may be tested in a phase I trial, often starting with the lowest dosage and escalating to higher dosages if tolerated. Phase II trials are designed to study the safety and efficacy of a new medication. These trials are often double-masked, in which neither patients nor investigators know what treatment is being administered. Classically these are called double-blind studies; however, in ophthalmology we prefer the term double-masked, because it is difficult to get a patient with an eye disease to enroll in a study with double-blind in the title. Phase II trials typically have more patients than phase I trials and are conducted in patients with the disease, but still may examine several dosages or treatment regimens. The phase III clinical trial is the pivotal clinical trial for the approval of the medication. These are almost always larger randomized clinical trials comparing the new medication to the standard treatment or to placebo. The US Food and Drug Administration (FDA) almost always requires two phase III trials with corroborative findings before approving a new drug. Studies conducted after a medicine is approved and marketed are called phase IV trials. These are conducted in patient populations for which the medicine is intended, and may compare the medicine to currently available therapies.
The randomized clinical trial provides the most powerful evidence about the value of a new therapy or diagnostic approach. Because patients enrolled into the study are randomly assigned into a specific group, if there are sufficient numbers of patients the groups are usually equivalent clinically. From an ethical standpoint, randomized clinical trials should compare treatments when the investigator is unsure which therapy is better. The situation where both treatment approaches have equivalent merits is termed clinical equipoise. ,
Even randomized controlled trials need to be designed appropriately and well conducted to ensure meaningful results. Key issues in the design and conduct of randomized clinical trials are listed in Box 6-1 . The investigators should have performed previous studies in the area and include people with appropriate training in the conduct and analysis of the study. The primary outcome of the study should be clearly stated, even if multiple outcomes are examined. The procedure for enrolling patients should be clearly delineated and state the inclusion and exclusion criteria. In addition, the therapy should be clearly stated, and the control group should be clinically appropriate. For example, if one were designing a trial to study the benefits of phacoemulsification as a technique for cataract surgery, it would be clinically inappropriate to compare this technique to intracapsular cataract extraction because extracapsular cataract surgery is the more frequently performed procedure.
