Feb 22, 2021 • 20 minutes read

How to get started with evidence-based optometry in your everyday practice

Introduction

Written by
Jon V.B. Gjelle

Busy clinical practices mean that we quickly get stuck in the groove – procedures and patients with common issues become routine. However, basing our clinical decisions on antiquated knowledge can have very dangerous implications for patient safety in our clinics.

Evidence-based optometry provides us with a framework to keep us up-to-date and examine the knowledge we use and the treatment and management recommendations we make. It also gives us tools to check the claims of newly developed products or management strategies.

In this handbook, I’ll advise you on how you can use evidence-based optometry in your everyday practice. I’ll start by covering some theory, such as the history of evidence-based medicine and why it’s important. Afterwards, I’ll give you some practical guides on how to find updated knowledge and how you can evaluate research for yourself – already from tomorrow and onwards.


Chapters

Five steps to make sure you keep up to date and continue to deliver the highest service quality to your patients

01
Do you know why it's called evidence-based?

How to connect with the roots of evidence-based practice and understand what evidence is.

Go to chapter
02
Is all knowledge created equally?

How to find good sources for evidence-based knowledge.

Go to chapter
03
Have you been using PICO?

How to formulate focused questions.

Go to chapter
04
I've found an article – what now?

How to appraise research and separate reliable from unreliable knowledge.

Go to chapter
05
Is it possible to incorporate evidence-based optometry in a busy practice?

How to keep improving on an everyday basis.

Go to chapter
01

Do you know why it's called evidence-based?

How to connect with the roots of evidence-based practice and understand what evidence is.

Although both medicine and other healthcare fields, in general, have a long history of research, implementing the knowledge acquired systematically through evidence-based practice is a relatively new development. Even more so for optometry, which is why we must start working towards getting it integrated into our daily work.

What is the story of evidence-based practice?

You’ve probably heard of “snake oil” – it’s an English term describing treatments or products that make false medical claims. It stemmed from the 1800s when quackery was common, with travelling salespersons going from town to town selling ointments and oils supposed to cure any illness, only to be gone the next day before people could realise that the products didn’t have any effect.

Although medicine has a long tradition of science and research, it wasn’t until the 1960s and 70s that medicine turned towards evidence-based practice. This shift was prompted by Alvan Feinstein’s Clinical Judgement and Archie Cochrane’s Effectiveness and Efficiency, documenting biases in the physician’s clinical reasoning and a lack of controlled trials supporting practices.

Evidence-based practice aims to incorporate knowledge from science and research into clinical decision-making to improve patient outcomes.

The concept was popularised by David Sackett from McMaster University and the Oxford Centre for Evidence-Based Medicine in the 80s and 90s. Although the movement began in medicine, several other healthcare fields, including optometry, has since started implementing evidence-based practice.

How do we evade biased research?

Evidence ranges from a low-to-high level of quality, depending on its inherent bias. The more bias research has, the lower its quality is.

At this point, you’re probably thinking: what is bias? Or you might have heard about bias before and learned that it’s a systematic research error. Simply put, bias is how something can be skewed to seem true or connected, but it only seems that way because the results are based on unfair data or analysis.

Bias is a genuine thing within humans, even outside science. If you discover information that plays along with your assumptions, you are more likely to believe that information is true even though it isn’t necessarily true.

So how do you get rid of bias? Since bias is inherent in humans, we cannot get rid of it altogether. But we can design our studies in such a way that bias becomes minimised. In short, anecdotal reports will have the highest bias. And systematic reviews with meta-analysis will have the lowest (see Figure 1: The Pyramid of Evidence).

In practice, we cannot only go off by systemic reviews. We also need everything from anecdotal reports to randomised-controlled trials to be able to do systematic reviews. Still, when implementing research into practice, we need to make sure that we rely heavily on systematic reviews since they have the highest level of evidence and lowest bias.

Figure 1: The Pyramid of Evidence

The patient’s perspective

One of the main criticisms of evidence-based medicine is that individual decisions about a patient are made based on population-level evidence without taking the patient’s own preference and perspective into account. Although evidence-based practice stresses the importance of research before making decisions, it shouldn’t rule out patient involvement. Evidence-based practice fits into person-centred care, as patients need to be informed of the possible courses of action there are in their specific case. This information needs to be evidence-based. In person-centred care, it should be the patient’s preference, after learning about their options and implications, that should guide how they receive care.

Future vision

There is also a valid concern that the evidence base in eye care is relatively small; there are generally few systematic reviews in eye care. A PubMed MeSH (Medical subject heading) term search for “eye” gives just over 350,000 results, but only 336 (0.10%) of those are systematic reviews. In comparison, a MESH term search for “heart” gives just under 510,000 results, and 800 (0.16%) of those are systematic reviews *.

Going from anecdotal evidence to systematic reviews is a slow process. It requires that many patients and clinicians participate in real-life research rather than having research limited only to academic populations.

Moving forward, it’s essential that we have a close collaboration between academics and clinicians, with more clinicians, clinician-scientists, and their patients participating in research in their daily practices.

* Search on PubMed conducted November 25, 2020, for terms “eye[mesh]” and “heart[mesh]” and then filtered for systematic reviews only.

02

Is all knowledge created equally?

How to find good sources for evidence-based knowledge.

Of course, knowing that evidence-based practice exists and why it’s important is only half the story. We need to know where we can find useful clinical information that can help us in our patient examination. And we need to be able to tell if we can trust the information we find or not.

Evidence-based guidelines and clinical decision tools

The best way for you to make sure that you’re aligned with evidence-based practice is to use evidence-based clinical practice guidelines. Multiple optometry associations have started producing them, but one of the premier resources for evidence-based practice guidelines is the American Optometric Association.

Since 2015, they have published three evidence-based clinical practice guidelines covering the adult eye and vision examination, the paediatric eye and vision examination, and care of patients with diabetes mellitus. Note that you should make sure your recommendations are aligned to national guidelines as well, even if they’re not evidence-based.

If you can’t find evidence-based guidelines, clinical decision tools are the next best thing. The British Medical Journal offers BMJ Best Practice – an evidence-based clinical decision tool covering several clinical topics. They have a separate section for ophthalmology which can be useful for optometrists as well.

Similarly to BMJ Best Practice, UpToDate is a clinical decision tool from Wolters Kluwer covering several medical fields, including topics related to ophthalmology and optometry such as amblyopia, strabismus, open-angle glaucoma, and age-related macular degeneration, among many others.

UpToDate

UpToDate is a clinical decision tool covering several medical fields, including topics related to ophthalmology and optometry.

Search in UpToDate

Another evidence-based clinical decision resource worth mentioning is Cochrane Clinical Answers. Cochrane Clinical Answers is a resource that answers specific clinical questions through data derived from published Cochrane Reviews, for example, “Does the use of spectacles or contact lenses slow the progression of myopia in children?” (view here).

Finding research

The issue with guidelines and clinical decision tools is that they take time to make. A certain number of systematic reviews need to be assembled in each area before guidance can be made. And if you need to answer a question that isn’t answered in a clinical decision tool already, you might need to look deeper into systematic reviews yourself.

Unfortunately, you will often find that there are few or no relevant systematic reviews – again because these also takes a long time to create and because the creation of systematic reviews is dependent on the existence of large amounts of lower-level knowledge. Therefore, as a clinician, you must sometimes cope with sources with a lower level of evidence. If this is the case, try to keep the pyramid of evidence in mind and stick to as high a level as possible (see previous Figure 1: The Pyramid of Evidence).

But where can you find research? The best source for locating research in medical sciences are two of the larger research search engines – PubMed and Embase.

PubMed is a US government-backed search engine that indexes the MEDLINE bibliographic database from the US National Library of Medicine and multiple journals. It’s free and available online here.

PubMed

PubMed is free to use and available online here.

Search in PubMed

Embase is Elsevier’s medical bibliographic database. Searching requires a subscription or access through an interface, such as Ovid. Embase also indexes MEDLINE, like PubMed, but includes additional European journals. Some national health libraries will provide access for practitioners in their country. Remember that what you are looking for might not be called “Embase” but might be called by an interface name, such as “Ovid.”

The road from idea to knowledge

Medical research often starts with a clinical question, observation, or intuitions in daily clinical practice. It can be anecdotal discussions between colleagues, a case report or case series. Questions will typically advance to be investigated through case-control and/or cohort studies. And later, if the question is treatment-related, it will be taken through randomised controlled trials. When enough randomised controlled trials exist, systematic reviews can be constructed through meta-analysis.

Please do remember this in your everyday clinical practice: health research is not inherently an academic thing – it is a clinical thing that starts with observations and questions you, as a clinician, have.

The process of scientific studies and knowledge delay

So, what is science, and how does it get published? Scientific studies must follow sound theoretical, ethical and methodological requirements. Medical studies, and studies with human subjects, typically have gone through ethical approval by a university or regional ethics board before researchers can start collecting data. Rules for how studies are designed, registered and started varies slightly by region and country.

Once a study is ongoing and researchers have collected data according to their methods, they get to the point of data analysis and writing up a discussion of their research. This will result in a finished draft article – called manuscript – but it can take a long time from the study starts until the manuscript is ready. Usually, researchers will have poster or paper presentations at scientific conferences before they finish the manuscript to get early feedback from peers and highlight new information that might influence their research.

Once the manuscript (draft paper) is written, researchers will submit it to a scientific journal for consideration. The editor of that journal can either reject it – typically if it doesn’t fit the journal’s scope – or send it to an editorial board member for peer review. The board member will be an expert on the concerning field and is responsible for selecting external reviewers and guiding the review process. A minimum of two scientists from the same field will sanity check the manuscript as external reviewers in peer review. Their task is to ensure the research is newsworthy and that there are no major faults within the manuscript, study design, methodology or analysis. Their input will help the board member form a suggestion to the editor to either accept, reject or revise the manuscript before a re-submission.

If the reviewers suggest a revision, and the editor agrees, the manuscript will go back to the authors to correct and address issues raised in peer review. After they’ve completed their revision, it must go through a new peer-review process, which can have the same outcomes as the first one. Once a manuscript is accepted, it will be published in an upcoming issue after going through copyediting and proofs.

Phew! It sounds like an expensive and time-consuming process, doesn’t it? According to a 2015 investigation of authors publishing in the field of conservation biology by Vivian M. Nguyen and colleagues, authors’ optimal review time was 6 weeks, compared to a typically experienced review time of 14 weeks.

And not only is it the review process that takes time. Data from 10,000 completed clinical trials registered on ClinicalTrials.gov with results published shows that the average time from a study is completed to results are published is just under 24 months! And the average time from its start date until results are published is just over 40 months *.

And this time will quickly add up when we need multiple studies to be carried out before a systematic review of the topic can be written. In other words, there can be a multiple-year knowledge delay between the time research questions arises and when we have a sound understanding of them. The time it takes to write and publish textbooks can further extend that delay. Therefore, although books are a great way to learn about new skills, they are not the tool to continuously keep updated on clinical knowledge. As a consequence of the delay, your knowledge might already be expired if you’ve not kept updated on the entire scope of your practice.

* Data collected on November 23, 2020, from 10,000 completed trials with results posted from ClinicalTrials.gov. The average time from the start of a study to the results are published is 40.4 months; the average time from completion of the study to results is 23.7 months.

Future vision

Optometry and ophthalmology need more systematic reviews and evidence-based knowledge for everything from examination methods to management and treatment. It will take years to get there, but the only way we can manage this is by having more clinicians participating in research, closer collaboration with academia, and more clinician-scientists actively working with patients in practice.

03

Have you been using PICO?

How to formulate focused questions.

In addition to knowing where we can find information, we also need to know how to ask the right questions. Using focused questions will make it easier for you to find relevant information suited for your patient.

How to structure your question using the PICO-model

The PICO model will help you to formulate a focused search question. PICO stands for Patient, Intervention, Comparison, Outcome. For each part of PICO, you will need to think through and define a specific group of patients, interventions, alternative treatments (comparisons), and which outcomes can be measured to say something about the effect of treatment (see below Table 1).

For example, we might be interested in finding out if orthokeratology is a preferred treatment over single-vision spectacles for teenagers with newly developed myopia. We can re-phrase this as a search question using PICO: In teenagers with newly developed myopia, would treatment with orthokeratology lenses in comparison to single-vision spectacles lead to slower progression of myopia and reduced risk of retinal disease?

Tabel 1: PICO

Going from a focused question to a search query

After you’ve got your focused question that encompasses all parts of PICO, set it up as a PICO table (see below Table 2). The PICO table has one row for each piece (Patient, Intervention, Comparison, Outcomes) and for each column in each row, you need to add relevant terms. For example, if we’re looking for ”teenagers with newly developed myopia”, we might think ”myopia” and maybe ”myopia AND teenager” are relevant searches for ”Patient”.

Table 2: PICO table for the focused (PICO) question ”In teenagers with newly developed myopia, would treatment with orthokeratology lenses in comparison to single-vision spectacles lead to slower progression of myopia and reduced risk of retinal disease?”

When we have the full PICO table, we can construct our search query. Between each PICO row, we add the operator “AND”, and between each term within one group, we add the operator “OR.” If a search term has multiple words, we need to add double quotation marks around them. We also need to add parenthesis around each PICO row. In our example, this might look something like this:

(myopia OR ”myopia AND teenager”) AND (orthokeratology OR ”corneal reshaping therapy”) AND (”single-vision spectacles” OR spectacles) AND (progression OR ”retinal disease”)

We’re now ready to use this search query in a bibliographical database such as PubMed/MEDLINE or Ovid/Embase.

Get it even more precise by using MeSH 

Another strategy that will give us more precise results is using Medical Subject Headings (MeSH).  MeSH is a controlled hierarchy of terms used by the National Library of Medicine to categorise and index PubMed/MEDLINE. This means that all articles on a specific subject will be covered by the same MeSH term, even if they don’t contain that word in the title or abstract. Each MeSH is categorised in a tree structure with relevant subcategories.

MeSH terms will limit your results and make them more relevant. If we take PubMed, for example, and you search for “eye,” you will get all articles that have “eye” in either its title, abstract or author/author affiliation. But search for the term “eye” in MeSH, and you will find results which have been categorised under the understanding of eye as: “The organ of sight constituting a pair of globular organs made up of a three-layered roughly spherical structure specialised for receiving and responding to light” (view here). Using MeSH on top of PICO will help you focus your search even further and get relevant results.

There are just two steps to include MeSH terms in your search; First, use the MeSH Browser to find applicable terms items in your PICO table. Secondly, to use MeSH terms in your search, you simply add “[MeSH]” immediately after your search term in PubMed, for example, “myopia[MeSH]” instead of “myopia.”

MeSH

Using MeSH on top of PICO will help you focus your search even further and get relevant results.

Search in MeSH

Be aware that newly added articles might not be MeSH indexed yet, so they might be omitted by the MeSH search. Embase has a similar system called Emtree. If you use Embase, you can read up on Emtree here.

Future vision

A great way to practise using PICO is by formulating focused questions and tables for as many patients as possible. Whenever you have time, try to think through your last examination:

  • What patient group did it involve, and what was the problem? (e.g., progressive myopia, convergence insufficiency, snow blindness)
  • What management did you prescribe, or have you heard of any management technique that would be the most often used alternative? (e.g., orthokeratology or single-vision spectacles, convergence training or prisms, no treatment, or cold compresses)
  • How can you tell if it has had any effect? (e.g., reduced progression of myopia, improved convergence, pain reduction).
04

I've found an article – what now?

How to appraise research and separate reliable from unreliable knowledge.

How do we interpret systematic reviews and research in general? When we’ve found a relevant research article, we need to appraise it to understand its strengths, weaknesses, and implications for our clinical practice. An appraisal should leave us with an idea of whether the article is valid for our patients and an interpretation of the results.

Four questions to ask yourself when evaluating research

According to the Centre for Evidence-Based Medicine at the University of Oxford, there are four questions you must ask yourself when evaluating research.

  1. Does the study address a focused question?
  2. Did the study use valid methods?
  3. Are the valid results relevant?
  4. Are the valid, important results applicable to your patient?

If your answer is “no” to any of these questions, you can stop reading and continue your search for other articles.

When you find an article where the answer is yes to all four questions, you should continue with one of the Centre for Evidence-Based Medicine’s appraisal tools. Most likely, you want to use the critical appraisal sheet for systematic reviews. The sheet gives you a specific set of questions that you should be able to answer while reading through the article you’ve found. It also includes information on where in the article to look for answers to those questions.

Find your answer in the forest plot

The most challenging question to answer is often, “What were the results?” This is because it involves making sense of forest plots in meta-analysis. A forest plot shows how each study included in meta-analysis relate to each other in terms of the odds ratio, relative risk or mean difference between groups.

As an example, look at Figure 3 from Walline and colleagues (2020). The figure shows a forest plot of the mean difference between patients treated with undercorrection versus full correction after one and two years of collecting data from three different studies.

For each study, there is a box with lines going out from it. The box’s relative size illustrates their considered weight in the analysis (higher weight means analysis favours the more reliable data from that study compared to less reliable data from other studies (this is typically reflective of how many participants the study had). The box is the point estimate of each study, and the lines on both sides of the box indicate a 95% confidence interval of the data. If the confidence interval crosses the central line at 0, it suggests that either treatment or placebo preference isn’t statistically significant.

The diamond below each comparison represents all studies’ combined mean difference and its 95% confidence interval. In this case, we see that after one year, the full correction has a mean difference of 0.15 D less than undercorrection.

Please note that due to low-level evidence, uncertainty, and the possibility that undercorrection accelerates eye growth, it is not recommended to undercorrect children as a strategy for myopia management.

Future vision

The Oxford Centre for Evidence-Based Medicine’s appraisal tools is an excellent resource, also for other types of research. Try it out whenever you hear of new research and discuss the appraisal with your local optometry department’s researchers and other peers.

05

Is it possible to incorporate evidence-based optometry in a busy practice?

How to keep improving on an everyday basis.

This handbook should have given you most of the tools and knowledge you need to start with evidence-based optometry in your practice. Let’s look at how we can implement evidence-based optometry in busy practices and how it can help you keep improving over time.

Implementing

Take a moment to think about what treatments you currently recommend and provide to different customers. How large is the evidence base of your recommendations? Set up PICO tables and conduct searches for your treatment strategies. Do the same the next time a salesperson introduces new products or treatments to you. Do they have adequate documentation to substantiate their claims?

Documenting

Another thing you should consider starting with is documenting what you do in your clinical practice. What types of diagnoses do you see, how are they managed, and how do they develop over time? This will give you great insight into your own clinical practice.

Ideally, your medical journal system should do this. If it doesn’t, this can also be tracked by a simple spreadsheet on your computer. Be mindful not to let the spreadsheet guide your decisions – they should still be based on the best available knowledge. A spreadsheet is merely a tool that gives you an overview of how you have been performing over time. And it also gives you the ability to check if you uncover the expected number of diseases compared to the demographic you’re working with.

Remember that the spreadsheet should be anonymised and not contain any specific patient information. I recommend you start with simple fields such as appointment date, appointment type (e.g., routine or follow-up), biological sex, year of birth, refraction, visual acuity, findings, treatment and time since the last appointment. It would also be wise to keep some sort of key to retrieve the patient’s journal, depending on what you are allowed to do in your jurisdiction.

Getting feedback

A database of what you do in clinical practice will also make it easier for you to start writing case reports on specific patients, thereby contributing to the greater knowledge pool that university researchers use when considering research questions for more extensive studies and reviews.

I strongly recommend you consider presenting interesting cases at conferences as well. This way, you get more direct feedback from your clinical peers, in addition to researchers.

Future vision

The more clinicians that contribute by documenting their clinical practice, writing and presenting case reports, and the more clinicians and universities that start to cooperate on treatment studies, the faster the systematic research in our field will grow, leading to better patient outcomes.