TrumpLyftAlles | 7 points | Sep 20 2020 03:08:17

ICTRP Search Portal (WHO's database of trials, 61 ivermectin trials on 2020-09-29)

https://apps.who.int/trialsearch/

permalink

[-] TrumpLyftAlles | 2 points | Sep 20 2020 03:29:52

This portal doesn't change its URL to reflect the search terms, so you'll have to do the search process to find all the ivermectin trials:

1) Enter "ivermectin AND covid-19" in the search box at the top and click the [Search] button

That will show you the first 10 search results, 10 being the default.

2) On the top-right corner of the search results page, select "100" from the "Show [dropdown] records per page" dropdown.

The page doesn't refresh, unfortunately.

3) Hit the browsers back arrow to return to the portal page and click [Search] again

This time the page will "remember" your preference and show you all 61 trials.

61 trials! {mind-blown emoji}

At the top of the results page are a couple [Export to...] buttons. I exported to CSV and opened the dump in Excel. My Excel skills are rudimentary, but I figured out how to filter on "Recruitment Status"

What do these means? My guess is "Not recruiting" corresponds to ClinicalTrials' "Active" or "Completed".

Hmm: Those add up to 59. Maybe a couple trials don't have "Recruitment Status" specified?

Whoops: WHO isn't complete! Filtering on "clinicaltrials.gov" leaves 37 trials, but ClinicalTrials actually has 39 trials when you search for "covid-19" and "ivermectin".

So there are at least 63 clinical trials out there.

63!! {mind-blown even more emoji}

Each WHO record contains a LOT of information. Dragging the exported CSV onto NotePad++, I could select the following as the data for one record:

"NCT04529525","14 September 2020","Ivermectin to Prevent Hospitalizations in COVID-19","Ivermectin to Prevent Hospitalizations in COVID-19: Randomized, Double-blind, Placebo-controlled","IVERCORCOVID19","Instituto de Cardiología de Corrientes","21/08/2020","20200821","9/20/2020 4:31:06 AM","ClinicalTrials.gov","https://clinicaltrials.gov/show/NCT04529525","Recruiting","No","18 Years","N/A","All","August 19, 2020","500","Interventional","Allocation: Randomized. Intervention model: Parallel Assignment. Primary purpose: Treatment. Masking: Quadruple (Participant, Care Provider, Investigator, Outcomes Assessor). ","Phase 2/Phase 3","Argentina"," ; ","Rodrigo Zoni, Dr;Rodrigo Zoni, Dr","","rodrizoni@yahoo.com.ar;rodrizoni@yahoo.com.ar","+5493794269233;+54379154269233","","
<br&rt;        Inclusion Criteria:
<br&rt;
<br&rt;          -  Over 18 years of age who reside in the province of Corrientes at the time of
<br&rt;             diagnosis;
<br&rt;
<br&rt;          -  Confirmed diagnosis of COVID-19 by polymerase chain reaction test for detection of
<br&rt;             SARS-CoV2 in the last 48 hours;
<br&rt;
<br&rt;          -  In the case of women of childbearing age, they must be using a contraceptive method of
<br&rt;             proven efficacy and safety (barrier, hormonal or permanent contraceptives) for at
<br&rt;             least 3 months prior to inclusion in the present study and for the entire period of
<br&rt;             time for the duration of the study and until at least 30 days after the end of this
<br&rt;             study. A woman will be considered to have no reproductive capacity if she is
<br&rt;             postmenopausal (at least 2 years without her menstrual cycles) or if she has undergone
<br&rt;             surgical sterilization (at least one month before the time of inviting her to
<br&rt;             participate in this study);
<br&rt;
<br&rt;          -  Weight at the time of inclusion greater than 48,000 kilograms;
<br&rt;
<br&rt;          -  That they sign the informed consent for participation in the study.
<br&rt;
<br&rt;        Exclusion Criteria:
<br&rt;
<br&rt;          -  Pregnant or breastfeeding women;
<br&rt;
<br&rt;          -  Known allergy to ivermectin or some of the components of ivermectin tablets or
<br&rt;             placebo;
<br&rt;
<br&rt;          -  Current use of home oxygen;
<br&rt;
<br&rt;          -  That require hospitalization due to COVID-19 at the time of diagnosis or history of
<br&rt;             hospitalization for COVID-19;
<br&rt;
<br&rt;          -  Presence of mal-absorptive syndrome;
<br&rt;
<br&rt;          -  Presence of any other concomitant acute infectious disease;
<br&rt;
<br&rt;          -  Known history of severe liver disease, for example liver cirrhosis;
<br&rt;
<br&rt;          -  Need or use of antiviral drugs at the time of admission for another viral pathology
<br&rt;             other than COVID-19;
<br&rt;
<br&rt;          -  Need or use of hydroxychloroquine or chloroquine;
<br&rt;
<br&rt;          -  Use of ivermectin up to 7 days prior to randomization;
<br&rt;
<br&rt;          -  Patients on dialysis or who have required it in the last 2 months or who plan to do it
<br&rt;             in the next 2 months;
<br&rt;
<br&rt;          -  Current participation or in the last 30 days in a research study that has included the
<br&rt;             administration of a drug.
<br&rt;      ","","Covid19","Drug: Ivermectin;Drug: Placebo","Percentage of Hospitalization of medical cause in patients with COVID-19 in each arm","","","","Yes","False","          ",""

Here's the same record without the inclusion and exclusion criteria, so you can operate the scrollbar while you're looking at the data:

"NCT04529525","14 September 2020","Ivermectin to Prevent Hospitalizations in COVID-19","Ivermectin to Prevent Hospitalizations in COVID-19: Randomized, Double-blind, Placebo-controlled","IVERCORCOVID19","Instituto de Cardiología de Corrientes","21/08/2020","20200821","9/20/2020 4:31:06 AM","ClinicalTrials.gov","https://clinicaltrials.gov/show/NCT04529525","Recruiting","No","18 Years","N/A","All","August 19, 2020","500","Interventional","Allocation: Randomized. Intervention model: Parallel Assignment. Primary purpose: Treatment. Masking: Quadruple (Participant, Care Provider, Investigator, Outcomes Assessor). ","Phase 2/Phase 3","Argentina"," ; ","Rodrigo Zoni, Dr;Rodrigo Zoni, Dr","","rodrizoni@yahoo.com.ar;rodrizoni@yahoo.com.ar","+5493794269233;+54379154269233","","
,"","Covid19","Drug: Ivermectin;Drug: Placebo","Percentage of Hospitalization of medical cause in patients with COVID-19 in each arm","","","","Yes","False","          ",""

OK, I expect y'all with Excel skills to go nuts. What cool insights can you come up with?

Edit:

I put the CSV file on Google Sheets. Would someone please confirm that you can open it? I'm a novice with Google docs (too).

The last two records are mostly missing data, even though it's present in the local CSV file, opened with Excel. Any guesses about why that would happen?

permalink

[-] TrumpLyftAlles | 2 points | Sep 20 2020 04:50:05

These are the columns on the ICTRP exported data. Unfortunately it has at least 2 defects:

1) The dose of ivermectin is not broken out into a column so you can sort and filter on it;

2) The study completion dates (primary and final) are missing.

I can probably fix the latter though code, for the ClinicalTrials trials, but for the other trials it will be a manual fix.

The former has to be manual, I think.

Do I have a volunteer? (seriously)

TrialID
Last Refreshed on
Public title
Scientific title
Acronym
Primary sponsor
Date registration
Date registration3
Export date
Source Register
web address
Recruitment Status
other records
Inclusion agemin
Inclusion agemax
Inclusion gender
Date enrollement
Target size
Study type
Study design
Phase
Countries
Contact Firstname
Contact Lastname
Contact Address
Contact Email
Contact Tel
Contact Affiliation
Inclusion Criteria
Exclusion Criteria
Condition
Intervention
Primary outcome
results Z posted
results date completed
results url link
Retrospective flag
Bridging flag truefalse
Bridged type
results yes no

permalink

[-] TrumpLyftAlles | 1 points | Sep 20 2020 04:17:01

Screen shot of the 61 trials on ICTRP with their Last Refreshed Date.

I'm really encouraged by the fact that 21 records were updated in September and 18 were updated in August. I take that as meaning those trials are actually happening.

Edit: Comparing the ICTRP Last Refreshed Date to the data on ClinicalTrials.gov for 3 trials, it appears that the Last Refresh Date is roughly the "Last Updated Posted" plus about 10 days.

Data is coming

permalink

[-] luisvel | 1 points | Sep 20 2020 17:25:33

This is very useful! I will try to make some time during the week and help with some

permalink

[-] TrumpLyftAlles | 1 points | Sep 21 2020 03:18:55

I wrote a long post about using ICTRP over on /r/covid19. I'll copy it to this topic.

In the meanwhile, I did a lot of exploring about how search works, and came up with the following improved search string:

covid19 and ivermectin OR covid-19 and ivermectin OR covid and ivermectin OR SARSCoV2 and ivermectin OR SARS-CoV-2 and ivermectin OR SARS-CoV2 and ivermectin OR sars-cov-2 and ivermectin

Result: 64 records found

There are at least 61 clinical trials underway, plus the three completed trials. :)

permalink

[-] TrumpLyftAlles | 1 points | Sep 21 2020 03:25:00

This site is how you search the World Health Organization's database of clinical trials. The list of data providers sending WHO their trials is on the portal page. It's one-stop shopping!

Until I found this yesterday, I thought there were 54 ivermectin vs covid-19 trials. This site turned up 61.. It's still not quite complete: the export file had 37 IVM trials from ClinicalTrials, where in fact there are 39. Looking at the "Last refreshed on" dates for a few CT trials, it looks like there's maybe 10-14 days lag, so trials added more recently than that won't be shown. That's a guess from a brief look at 3-4 trials.

Here it explains that it handles trials that may appear in multiple sources:

To reduce confusion and as a service to aid in the unambiguous identification of trials, the ICTRP Search Portal attempts to bridge (group together) such multiple records about the same trial. It does this by matching main trial identifiers to secondary trial identifiers found in trial records. The ICTRP Search Portal shows that it has found a group of records about the same trial by displaying a plus (+) sign in its search results:

You can download the entire database in CSV or XML format from this page.

Searching

Tip: Read the search tips page.

On that page it says:

The standard search looks for terms (words or phrases) in Title, Primary sponsor, Health Condition(s), Intervention(s), Countries of recruitment, Main ID, Secondary ID(s)

Use "and" to separate search terms, e.g.

"ivermectin and covid-19"

You can put two "and" statements between an "or":

"covid19 and ivermectin OR covid-19 and ivermectin"

The case of the "and" and "or" doesn't matter: "AND" works just like "and".

When you search on the portal page, the results page will show you only the first 10 trials. In the top-right corner of the results page is a "Show [dropdown] records per page" dropdown. If you select 100, the page doesn't refresh; you have to go back to the portal page using the [Back to Search] button on the top left. Then you'll see the first 100 records. It may take you a couple tries (it's flaky).

The results page only has a few columns. A useful one is the rightmost "Results available" column, if you're looking for completed studies. AFAIK there's no way to filter the trials by status, e.g. recruiting or completed. You'll have to export the data and filter in Excel or your tool of choice.

You click the study title on the results page to go its detail page.

Search bugs

The tooltip quoted above says that search looks at "Countries of recruitment". ClinicalTrials is smart about this: I can search for condition "covid-19" and put "ivermectin egypt" in the "other terms" box, and CT will find the 6 trials being done in Egypt.

That doesn't work with the WHO portal. Searching for "covid-19 and ivermectin and egypt" returned 5 trials, which sounds like it's in the ball park -- but at least one of them had no mention of "ivermectin" anywhere.

On the search tips page, it says:

Synonyms: The standard search looks for trial records containing synonyms of search words or phrases using the UMLS metathesaurus

Here is a UMLS FAQ that I didn't find helpful.

Clinical Trials is intelligent: covid-19, covid19, sars-cov2 and sar-cov-2 all return the same records.

Sarscov2 returns 0 records on CT.

This is now many record are returned from the WHO database for each term:

SARSCoV2 5
SARS-CoV-2 1209
SARS-CoV2 299
Covid-19 5749
Covid19 814

It appears that there are no synonyms being applied.

"covid-19 and ivermectin" returns 61 records.
"covid19 and ivermectin" returns 13.

These two studies were found among the 13:

These were not among the 61 records because "covid-19" appears nowhere in the study, not in the title or in any of the searched details.

This lack of support for synonyms makes searching a nightmare.

"SARS-CoV-2 and ivermectin" returns 8 records, 1 of which was not among the 61 and wasn't one of two found by the other search. There are 64 studies of interest.

How do you get them all in one result set?

The obvious solution to the problem that "covid19 and ivermectin" returns records that "covid-19 and ivermectin" would be to apply logical operators. On the search tips page it say:

Combine term searches using Boolean operators: * OR: find any of the terms * AND: find all the terms * NOT: find the first term without the second term

Looking for both "covid19" and "covid-19" in one search:

"covid19 and ivermectin OR covid-19 and ivermectin"

returns 63 records, 61 plus the 2 from the 13 records that were not part of the 61. Yay!

Throwing in the of the SARS terms:

"covid19 and ivermectin OR covid-19 and ivermectin OR sars-cov-2 and ivermectin" returns 64 records!

Other bugs

There are pagination buttons at the top and bottom of the results page that look like "1 2 3 4 5 6 7 8 9 10 _ " where each item has an underscore. Clicking them doesn't do anything. I tried in Chrome, FireFox and Edge: no go.

AFAIK, there's no way to see the 101th trial. If you have more than 100 in your result set, your only option is to export them to CSV or XML.

I wanted to see if all the trials would be available when exported. I did two experiments, exporting result sets of 299 and 243 records to CSV and opening the file in Excel. In both cases, I only got 215 trial records.

I tried exporting to XML. Nothing was exported, esp. no file appeared in my browser download folder.

If your trial search returns more that 215 records, you'll need to break it up into smaller result sets. Try applying the NOT like "ivermectin and covid-19" and "ivermectin not covid-19" to get all the ivermectin trials without passing 215 records per result. The Phases in the listbox under where you enter the search string might help, depending on your data; they do change the result set. I didn't test whether you get all the records, if you successively try all the phases.

Click. Nothing.

That makes sense, since from all appearances, that's the default behavior.

I can do web development. Will WHO let me code from home? {half-kidding emoji}

What data is missing?

The database does not include completion dates, the Estimated Primary Completion Date and Estimated Study Completion Date that appear on ClinicalTrials.gov. The detail page has a link to the trial page, so if it's a CT trial, you can easily go there for those dates. The completion dates are probably omitted because some registries don't have them, e.g. the EU registry.

It does have the "Date of first enrollment", which seems to be CT's "First Posted" date, plus lag (brief look). Whoops, found a counter-example. I don't know what it is.

Secondary outcomes are not included in the WHO database, so you'll have to go to CT (or whichever registration site) for those.

Some fields are missing data, e.g. the Contact Email is there about half the time.

I was disappointed that dosages are not broken out into their own column, so I could find the studies dosing at 600mcg/kg, for example. Instead you get the whole treatment line as a text string. It would be a lot of work to break out the dosages separately (which I'll be doing). CT doesn't have the dosages broken out, but it supports searching for any string, so it's easy to find the trials testing 600mcg/kg (might get some false positives).

What features are missing?*

ClinicalTrials has a bunch of filters on the right side of the results page than let you select a subset of the returned records. "Have any studies been completed since I checked last"? Check the "Completed" checkbox in the Recruitment list and hit the [Apply] button -- and you're only looking at completed studies.

Neither the WHO portal or ClinicalTrials have sort capability, which is a shame.

Exporting data

At the top of the results page are buttons to click to download the results in CSV or XML format. (XML export doesn't work: click the button, an agreement dialog pups up, you say OK, and then -- no data is exported to the browser download folder.

It appears that the exported data is the same as that shown on the details page. The full text of the inclusion and exclusion criteria is formatted with <br tags so it will look like ordinary text on an HTML page, with the source's line spacing preserved.

Given the search, display of search results, and export problems -- it might be easier to export the entire WHO database, bring it into Excel, and use SQL, if that's familiar. Or Excel's sort and filter features; I don't know much about those.

TL;DR: This portal to the WHO trial database is an easy way to find all (or nearly all?) the clinical trials matching your search criteria. It may not have the newest trials. Search is flawed, esp. it doesn't do synonyms for handle extended logical expressions. It won't won't display more than 100 records. Data can be exported to CSV or XML format.

permalink

[-] TrumpLyftAlles | 1 points | Sep 21 2020 03:46:35

Export XML works! I just didn't know how to work it.

On the result set page, it says:

3407 records for 3145 trials found for: vitamin d

I think it means that about 260 trials were duplicated across registries and merged: 3407 were merged into 3145 trials.

Click [Export results in XML] on the top right of the results page.

Click [I agree] on the Terms and Conditions page that pops up.

Wait a while.

When the page refreshes, there will be two new buttons. On the left, [Export selected records to XML]. You check the checkboxes to the left of the studies you want to export, then click the button.

[Export all trials to XML] exports them all! I checked: it works. 3145 exported. :)

permalink

[-] TrumpLyftAlles | 1 points | Sep 21 2020 03:54:29

Above, I figured out a search that returns 64 ivermectin vs covid-19 studies. The 64 doesn't include this one (might be too know). So 65!

Oh yeah: I forgot that WHO only has 37 from ClinicalTrials when there are 39, so it's missing 2. 67!

permalink

[-] TrumpLyftAlles | 1 points | Oct 01 2020 21:36:23

This is frustrating: The WHO data doesn't include the complete treatment information, or doesn't for at least this trial: NCT04351347.

ICTRP has:

Combination Product: Ivermectin plus Nitazoxanide
Other: Standard Care

ClinicalTrials has:

Experimental: Ivermectin plus Nitazoxanide
Ivermectin 200 mcg/kg once orally on empty stomach plus
Nitazoxanide 500 mg twice daily orally with meal for 6 days

I wanted to get all of the doses of ivermectin across the 65 trials. I thought I could just get it from the WHO data.

SHIT.

permalink