To assist with the interpretation of results of research, all trials indexed in PEDro are assessed for methodological quality. Trials are ranked by quality criteria so users can have rapid access to the most valid trials. The PEDro scale is used to assess the methodological quality of trials (Appendix 1) and it is comprised of 11 items. The first item relates to external validity. The remaining 10 items are used to calculate the final score, which ranges from 0 to 10. The purpose of the PEDro score is to help users identify trials that have good internal validity (items 2–9) and that report enough data to make their results interpretable (items 10–11). The total PEDro score as well as the score for each item is available for every trial indexed in PEDro. This unique feature is crucial for facilitating the use of high-quality clinical research to inform clinical practice, helping users to overcome the barrier of lack of time and skills in critical appraisal. Systematic reviews and clinical practice guidelines are not rated.

Several studies have demonstrated the reliability and validity of the PEDro scale.5–8 The PEDro scale has been translated and cross-culturally adapted into Brazilian Portuguese following best practice guidelines, so that clinicians who are not proficient in English can also use the scale (Appendix 2). The Portuguese version has acceptably high reliability, similar to the original English version.9

PEDro indexes articles that report randomized controlled trials that investigate the effects of physical therapy interventions, and systematic reviews and evidence-based clinical practice guidelines that are based around such trials. All articles that are indexed in PEDro must be peer-reviewed and published as full-text manuscripts. PEDro indexes any articles that meet these criteria, with no restriction by language, journal of publication, or year of publication. For each indexed article, PEDro archives the bibliographic details, the abstract, and links to full-text versions of the article that are available elsewhere online. The specific criteria for trials, reviews, and guidelines to be included in PEDro can be found in Appendix 3.

The articles indexed in PEDro are identified via comprehensive search methods, including automated optimized searches conducted in bibliographic databases (Medline, Embase, CINAHL, PsycINFO, AMED, Cochrane Database of Systematic Reviews, Cochrane CENTRAL Register of Controlled Trials) as well as web-sites of clinical practice guidelines (e.g., https://www.nice.org.uk/guidance, https://www.sign.ac.uk/our-guidelines.html). Citation tracking is also performed in systematic reviews included in the PEDro database. Therefore, users only need to access a single database to search for answers for questions about the effects of physical therapy interventions.

PEDro is the most comprehensive database of trials, reviews, and guidelines in physical therapy. We have conducted an analysis of the contents of PEDro and extracted data available up to August 2019. There was a total of 44,309 articles indexed in PEDro, including 34,619 randomized controlled trials (78%), 9004 systematic reviews (20%), and 686 clinical practice guidelines (2%). Fig. 1 shows the cumulative number of trials, reviews, and guidelines available each year.

Fig. 1.

Cumulative number of articles reporting randomized controlled trials, systematic reviews, and evidence-based clinical practice guidelines indexed in PEDro each year.

(0.07MB).

The oldest trial indexed in PEDro was published in 1929. Since then, there has been exponential growth of the number of trials in physical therapy, with 560 trials published in 1999 (the year PEDro was launched) and 2104 in 2018. We used an exponential growth curve to fit the data, resulting in a very good fit (number of trials=e 0.12 x years since 1929; r2=0.97). The number of articles reporting trials is doubling about every 6 years. Based on this modelling we estimate that PEDro could contain 70,000 trials in 2025.

The first archived systematic review and clinical practice guideline were published in 1982 and 1987, respectively. The number of systematic reviews and clinical practice guidelines has increased rapidly, but at a slower rate than for trials. The number of eligible reviews that are published annually has increased from 68 in 1999 to 858 in 2018. The data are similar for practice guidelines, increasing from 3 in 2008 (when guidelines were first indexed in PEDro) to 46 in 2018. Like for trials, exponential growth curves fit these data well: number of reviews=e 0.25 x years since 1982 (r2=0.85); number of guidelines=e 0.21 x years since 1987 (r2=0.90). The number of articles reporting systematic reviews and clinical practice guidelines is doubling about every 3 years. Based on this modelling we estimate that PEDro could contain 36,000 reviews and 2800 guidelines in 2025.

Articles indexed in PEDro were published in 30 languages. The vast majority of articles were published in English (90%), followed by Chinese (5%) and German (1%). In August 2019, 219 articles in Portuguese (130 trials and 89 systematic reviews) were indexed in PEDro.

The articles indexed in PEDro cover all areas of clinical physical therapy practice and are categorised into 10 subdisciplines (Fig. 2). Each article can be coded for up to three subdisciplines of physical therapy. Articles are coded as “other” when it is not possible to code it using the available subdisciplines. Since PEDro was launched, one subdiscipline (oncology) was added in 2002 in response to the growing body of research about physical therapy interventions for this clinical population. The largest proportion of articles indexed in PEDro are in the musculoskeletal subdiscipline, followed by cardiothoracics (Fig. 2).

Fig. 2.

Number of articles reporting randomized controlled trials, systematic reviews, and evidence-based clinical practice guidelines according to the subdisciplines of physical therapy. Note: as each article can be classified for more than one subdiscipline, the total number of articles in this graph does not match the total number of articles in PEDro.

(0.13MB).

Each trial in the PEDro database is appraised for methodological quality by two independent raters using the PEDro scale. The percentage of trials indexed in PEDro fulfilling each item of the PEDro scale is shown in Fig. 3A. Most trials employ random allocation (97%), report between-group comparisons (94%), and provide point estimates and measures of variability of data (91%). Trials infrequently report: blinding of therapists (1%) or subjects (6%); concealed allocation (26%); and, intention to treat analysis (27%). The mean PEDro score in August 2019 was 5.1 out of 10 (standard deviation, 1.5). Approximately one third of trials (37%) had a PEDro score equal to or greater than 6, and therefore are considered to be of moderate to high quality. Overall the quality of the trials in physical therapy has increased since 1929 (Fig. 3B).

Fig. 3.

A Percentage of articles reporting randomized controlled trials indexed in PEDro that satisfy each item of the PEDro scale. B. Total PEDro score (mean and standard deviation) of articles reporting randomized controlled trials indexed in PEDro since 1929, displayed in 5-year intervals (except for the first and last intervals, which are 20 and 6 years, respectively). The number of trials for each period is indicated at the top of each bar. Note: this graph is based on ratings for 33,787 trials with complete data. The remaining 832 trials indexed in PEDro are in-process, so have not been rated using the PEDro scale yet.

(0.15MB).

Over the 12 months between 1 July 2018 and 30 June 2019 the PEDro database was widely used, with users from 214 countries (Fig. 4A). In the period investigated, 2,890,630 searches were performed, meaning that 5 new searches were initiated every minute. Brazil was the leading user of PEDro, accounting for nearly one quarter of the searches (675,533, 23%). The other countries with high usage were United States of America (9%), Spain (8%), Australia (7%), and France (5%).

Fig. 4.

PEDro searches across the world (A) and in Brazil (B) for the year between 1 July 2018 and 30 June 2019. Darker colour indicates greater number of PEDro searches.

(0.19MB).

PEDro was used across all states in Brazil, with the highest usage rates concentrated in the southeast area (Fig. 4B). The states that mostly commonly used PEDro were: São Paulo (29%), Minas Gerais (15%), and Bahia, Ceará, and Rio de Janeiro (all 6%). These five states together accounted for 62% of PEDro usage in Brazil.

Because PEDro only indexes trials, reviews, and guidelines related to physical therapy, it can be used to search efficiently for high-quality research to answer clinical questions about the effects of physical therapy interventions. Users can search PEDro with either of two interfaces. Simple searches can be conducted using one interface (http://search.pedro.org.au/search), which consists of one field where users can enter free text. The database then returns all articles for which all of the entered text is included in the article’s title or abstract. More advanced searches can be conducted using the other interface (http://search.pedro.org.au/advanced-search). Here, search criteria can be entered into any of 13 optional search fields. Six of the fields (Therapy, Problem, Body Part, Subdiscipline, Topic, and Method) have pull-down menus with options that correspond to the coding of articles in the database. The remaining fields allow free text to be entered: Title & Abstract, Title Only, Author/Association, Source, Published Since, New Records Added Since, and Score of at least. The interfaces are shown in Figs. 5A and 5B. As a general guide, physical therapists should use the advanced search interface as this encourages greater precision when entering search terms to answer clinical questions.10,11

Fig. 5.

The simple (panel A) and advanced (panel B) search interfaces for PEDro.

(0.13MB).

While both interfaces are intuitive, the searches entered by users have been analysed and some common errors were identified.10 For example, Boolean operators and parentheses do not function within any individual search fields in the advanced or simple search interfaces. Also, bibliographic details of articles are entered into PEDro using only American Standard Code for Information Interchange (ASCII) characters. Therefore, users who are searching with a name such as “Gonçalves” in the Author/Association field or the term “esclerose múltipla” in the Title Only field would have to enter “Goncalves” or “esclerose multipla” to conduct a valid search. Therefore, automated alert messages were added to the interface to notify readers when they try to enter erroneous text in one of the free-text search fields. These alerts have led to a small reduction in the number of erroneous searches.11

When a search is conducted, the results of the search are presented in a particular order: guidelines are shown first (sorted by year), followed by reviews (with reviews published in the Cochrane Database of Systematic Reviews appearing before reviews published in other journals, both sorted by year), and then trials (from highest quality to lowest quality). This appears to encourage users to click on more synthesized and higher-quality research within the search results.12

Searching PEDro is free for anyone who has access to the internet. Almost all articles in PEDro show the abstract and about half of the articles in PEDro also have links to free full text online, so even users who do not have access to a library of journals will still find it fruitful.

A service was added to PEDro in 2015 to assist users who would like to regularly browse the latest content. This service is called ‘Evidence in your inbox’. Users can sign up to receive free monthly updates via email that detail the most recent additions to PEDro in up to 15 areas of practice. This has been a very popular service, with over 13,600 subscribers in August 2019. Brazil has a high rate of uptake, accounting for 35% of subscribers. Engagement rates for all areas exceed the industry average for number of opens (up to 29% for ‘cerebral palsy’), and click-throughs (up to 8% for ‘musculoskeletal’). Clinicians, researchers, and educators value the service very highly, using the feeds for journal clubs, to inform practice, and to prepare lectures.

Skills in evidence-based physical therapy include being able to ask a clinical question, find relevant clinical research, appraise the quality of the identified research, and integrate the findings into clinical decision making for individual patients. Many features have been implemented within and alongside the original database to assist users to develop these skills.

To facilitate the use of PEDro we have made the web-site and YouTube channel available in 13 languages.13 Both the web-site and YouTube channel are available in Portuguese, English, Simplified Chinese, Traditional Chinese, French, German, Italian, Japanese, Korean, Spanish, and Tamil. The web-site is also available in Turkish and the YouTube channel in Dutch. Among these, the English-language sections of the web-site had the largest number of pageviews (83%) followed by Portuguese (11%). The most commonly viewed languages in the PEDro YouTube channel were English (59%) and Portuguese (20%). We also provide newsletters, blogs, and social media (Facebook and Twitter) in English and Portuguese language.

PEDro has developed tutorials, including videos, to support physical therapists in developing their searching and appraisal skills. We have published several “How to” videos on the PEDro web-site including: how to ask a clinical question, and how to perform a PEDro simple or advanced search. The videos can be accessed via the PEDro search help pages (http://www.pedro.org.au/portuguese/search-help/) and the PEDro YouTube channel (https://www.youtube.com/c/PhysiotherapyEvidenceDatabasePEDro; see the Português playlist).

The PEDro scale has been developed for rating the validity of trials through critical appraisal. To facilitate the use of the PEDro scale, we have launched a training program for PEDro raters. It provides a definition and detailed description of each item on the scale, relevant references, answers to frequently asked questions, examples, and practice articles for each item of the PEDro scale. The PEDro scale training program is currently available in English and Portuguese. The training is online through a 3-month subscription for a small fee (https://training.pedro.org.au/).

Using high-quality clinical research to inform practice is a professional imperative. This will require deliberate and sustained efforts from individual physical therapists and the organizations and businesses that support physical therapists to practice safely and effectively. Individual physical therapists need to engage in life-long learning to develop their skills in areas like evidence-based physical therapy (including posing questions for clinical uncertainty, searching to identify high-quality clinical research to answer these questions, reading research (including critical appraisal), and implementing changes to their practice). Organizations and businesses can strive to produce, report, and promote the best research to guide practice. There are several ways that this could occur. Educators could ensure that training programs are evidence-based and include strategies for changing practice as new knowledge is developed. Researchers, research funders, ethics committees, and journals could ensure that research to answer important questions is conducted rigorously and clearly communicated (it would be nice not to have to wait until about 2056 for all new physical therapy trials to achieve a total PEDro score of 8/1014). Funders of physical therapy services could reward the implementation of evidence-based care. Peak physical therapy groups like the World Confederation for Physical Therapy and its member organizations could promote the best research to their members and support local and global initiatives like PEDro.15

PEDro will continue to be developed to facilitate the use of high-quality clinical research by physical therapy clinicians, educators, students, and researchers. In October 2019 PEDro celebrated its twentieth anniversary. Some key enhancements to mark this important achievement include:

• 1

  The PEDro web-site and training materials, already available in 13 languages, will be expanded to include Arabic and Ukrainian.

• 2

  Ensuring that the content distribution platforms used for PEDro are accessible world-wide (including in countries like China).

• 3

  Adding new subdiscipline codes (and associated ‘Evidence in your inbox’ feeds) in response to the evolution of physical therapy practice. Codes that are currently under consideration are for mental health and metabolic conditions (including diabetes and obesity).

• 4

  Launching a new database called DiTA (Diagnostic Test Accuracy) that focuses on the accuracy of diagnostic tests used by physical therapists.16

PEDro will continue to refine and initiate strategies like ‘Evidence in your inbox’ to promote the results of high-quality clinical research that are of high importance to the global physical therapy community. Because PEDro is produced on a not-for-profit basis with support from industry partners, these initiatives will be dependent on attracting sufficient funding.