In this post, we present research on substance use in a COVID-19 setting. The results are based on publications registered in PubMed as published from January 1^st, 2020, to June 5^th, 2021.

The analysis of all research including covid-19 terms, AND substance use-related terms in title or abstract in PubMed from January 1^st, 2020 to June 5^th, 2021, is based on 2243 articles, which gave 510 keywords that appear in 5 or more manuscripts, that divided into 5 clusters.

The red cluster could be called “risk factors.” It includes four main topics: a) comorbidity, underlying diseases, hospitalization, and treatment.  b)  age factors and how taste and smell may be affected. C) risk assessments and time factors, and d) a small cluster on pain and pain management.

The green cluster could be called “substance use disorders.” This cluster includes two main topics: a) opioid-related disorders and treatment of this with, e.g., methadone and buprenorphine. Telemedicine, prevention, and harm reduction are also central to this topic. b) Another topic is public health.

The yellow cluster could be called “smoking.” It includes one main topic: research on smoking nicotine, e-cigarettes, lungs, and inflammation.

The purple cluster could be called “mental health and health behaviors.” It includes two main topics: a) research on mental health, including anxiety and depression, stress, and psychological adaptation. b) A second cluster includes research on quarantine and lockdown, social isolation and loneliness, and different lifestyle factors and health behaviors such as exercise, sleep, alcohol use, diet.

The blue cluster could be called “alcohol and infection control.” The cluster includes four main topics: a) epidemiology and infectious diseases, outbreaks and testing, b) research on preventative measures such as the use of hand sanitizers and masks. c) A third topic is on hospitals, including both health care personnel and patients. d)  alcohol in relation to infectious disease control and public policy.

The search: title/abstracts that included any of the following:

• COVID-19,
• SARS-CoV-2,
• severe acute respiratory syndrome
• coronavirus 2,
• 2019 novel coronavirus
• 2019-nCoVAND any of the following:
  

  alcohol, tobacco, smoking, snus, opioid, opiate, methadone, buprenorphine, cannabis, amphetamine, methamphetamine,

  nicotine, cocaine, ecstasy, MDMA, hash, heroin, kratom, hallucinogen, GHB, inhalants, ketamine, marijuana, khat, synthetic cathinone,

  benzodiazepine, z-hypnotic, cannabinol, LSD, snuff, pcp, thc, anabolic steroid, narcotic, depressants, illicit substance, illegal substance, psychoactive drug.

In the table below, we have listed the 20 most central keywords in each cluster.

To open the map directly in VOSviewer, please use the Keyword co-occurrence MAP and NET files on CovidSubstanceUse saved in June 2021, available through the Open Science Foundation (OSF) repository: https://osf.io/54gqw/

In this post, we present the current focus of the COVID-19 literature, based on publications registered in PubMed as published from April 5th to May 2nd, as well as all articles published up to this date.

The analysis of all research in PubMed up to 2nd May, is based on 125 521 pulications, which gave 4160 keywords that appear in 10 or more manuscripts, that divide into 6 cluster.

The first four clusters are reasonably similar in size, while cluster 5 is about 60% of the size of cluster 1, and cluster six only has 24 terms.

For the month of April 2021:

During this time, there were 8976 publications, and 503 keywords occurred in ten or more manuscripts. The keywords divided into 5 clusters. To create the map, we used the LinLog method.

In the table below, we have listed the 10 most central keywords in each cluster.

To open the map directly in VOSviewer, please use the Keyword co-occurrence MAP and NET files from March, available through the Open Science Foundation (OSF) repository: https://osf.io/54gqw/

In this post, we present the current focus of the COVID-19 literature, based on publications registered in PubMed as published from March 8th to April 4th. During this time, there were 9653 publications, and 637 keywords occurred in ten or more manuscripts. The keywords divided into 5 clusters. To create the map, we used the LinLog method.

In the table below, we have listed the 10 most central keywords in each cluster.

To open the map directly in VOSviewer, please use the Keyword co-occurrence MAP and NET files from March, available through the Open Science Foundation (OSF) repository: https://osf.io/54gqw/

In this post, we present the current focus of the COVID-19 literature, based on publications registered in PubMed as published from February 8th to March 7th. During this time, there were 9741 publications, and 654 keywords occurred in ten or more manuscripts. The keywords divided into 7 clusters. To create the map, we used the LinLog method.

In the table below, we have listed the 10 most central keywords in each cluster.

To open the map directly in VOSviewer, please use the Keyword co-occurrence MAP and NET files from March, available through the Open Science Foundation (OSF) repository: https://osf.io/54gqw/

In this post, we present the current focus of the COVID-19 literature, based on publications registered in PubMed as published during January 2021. During this month, there were 9861 publications, and 1239 keywords occurred in five or more manuscripts. We used two resolutions this month, both 1, and 1.5. The keywords divided into 6 and 12 clusters respectively. To create the map, we used the LinLog method.

In the table below, we have listed the 15 most central keywords in each cluster in the higher resolution map.

The topic in each cluster seems to be:

Keyword Co-occurrence map: resolution 1.0

Keyword Co-occurrence map: resolution 1.5

To open the map directly in VOSviewer, please use the Keyword co-occurrence MAP and NET files from January 6th, available through the Open Science Foundation (OSF) repository: https://osf.io/54gqw/

In this post, we present the current focus of the COVID-19 literature, based on publications registered in PubMed as published during December 2020. During this month, there were 7 148 publications, and 907 keywords occurred in five or more manuscripts. We used a higher resolution this month, 1.5, and then the keywords divided into 10 clusters. To create the map, we used the LinLog method.

In the table below, we have listed the 15 most central keywords in each cluster. The topic in each cluster seems to be:

• Cluster 1: Vaccines.
• Cluster 2: Health care delivery
• Cluster 3: Mental health issues
• Cluster 4: Treatment
• Cluster 5: Public health
• Cluster 6: Testing
• Cluster 7: Risk factors and mortality
• Cluster 8: Children and socioeconomic factors
• Cluster 9: Pregnancy
• Cluster 10: Demographic

To open the map directly in VOSviewer, please use the Keyword co-occurrence MAP and NET files from January 6th, available through the Open Science Foundation (OSF) repository: https://osf.io/54gqw/

In this post, we present the current focus of the COVID-19 literature, based on publications registered in PubMed as published during October and November. During these two months, there were 21 758 publications, and 649 keywords occurred in eight or more manuscripts. These divided into four main clusters. To create the map, we used the LinLog method. The minimum cluster size was set to 10.

To open the map directly in VOSviewer, please use the Keyword co-occurrence MAP and NET files from December 1st, available through the Open Science Foundation (OSF) repository: https://osf.io/54gqw/

In this week’s update, we present the current focus of the COVID-19 literature, based on publications registered in PubMed as published during the past three weeks (September 28th to October 18th). There were 527 keywords that occurred in five or more manuscripts. These divided into five main clusters. To create the map, we used the LinLog method. The minimum cluster size was set to 10.

To open the map directly in VOSviewer, please use the Keyword co-occurrence MAP and NET files from October 19th, available through the Open Science Foundation (OSF) repository:
https://osf.io/54gqw/

Corpus downloaded from PubMed September 28th consisted of 53994 publications. There were 3300 keywords that occurred in five or more manuscripts. The keywords divided into five main clusters. To create the map, we used the LinLog method. The minimum cluster size was set to 30.

Table of the most central keywords in each cluster

(Centrality measure used is weighted degree centrality/total link strength)

To open the map directly in VOSviewer, please use these the latest MAP and NET files, available through the Open Science Foundation (OSF) repository:
https://osf.io/54gqw/ 

In this post, we compare the focus of the COVID-19 literature published in the very early stages of the pandemic (up to March 23rd) to the focus of literature published during August 2020.

The first graph below is based on the literature based on the Corpus downloaded from PubMed 23rd March. The literature consisted of 1323 publications, of which there were 225 keywords that occurred in three or more manuscripts. These divided into four clusters, that are described in this preprint

The second graph below is based on COVID-19 literature published during August 2020. The literature consisted of 10901 publications, of which there were 1179 keywords that occurred in 5 or more manuscripts. The keywords divided into five main clusters. To create the map, we used the LinLog method. The minimum cluster size was set to 15. We provide a brief description of the clusters from the August COVID-19 litterature: Within cluster 1 (red), the “Pandemic of viral pneumonia; public health and world spread”, we found topics relating to “Public health concerns” “health personnel, education and behavior”,  “Guidelines and health care”, and “vulnerabilities in health care”. Cluster 2 (green), “Treatment of COVID-19”, include topics relating to “Epidemiology of virus transmission” “Antiviral agents and vaccines”, “Cancer and chemotherapy”, “Angiotensin and diabetes” and “Antibodies and passive immunization” Central themes in cluster 3 (blue), “Epidemiology of the pandemic and disease”, are “Respiratory illness and care”, “Risk factors and mortality”, “Clinical investigation and diagnosis” and “Population epidemiology”. Central themes in cluster 4 (yellow), “Surveys of disease and mental health”, are topics relating to “Population surveys and anxieties” and “Health care and migration”. Cluster 5 (purple), is dominated by one single theme, namely topics relating to “Lab-work and diagnostics”; which is also what we labeled this cluster.

When inspecting and comparing the network graphs illustrating the COVID-19 literature published during August and one published during the early phase (January 1^st – March 23^rd, 2020), we notice some consistent themes, and some emerging themes. Much of what was being covered during the early phase is still in focus. For instance, the topics from the “health and pandemic management” cluster from the early phase is reflected in the topics covered in the “Pandemic of viral pneumonia; public health and world spread” cluster in the August literature network graph. Early in the pandemic, there was not however, a focus on mental health issues. This has become an additional focus later in the pandemic, and is evident in the “Surveys of disease and mental health” in the August literature network graph. These are just some examples of insights that can be gained from comparing network graphs from different time periods.

Figure 1: The research focus in the very early Covid-19 literature

Figure 1 Notes: Red cluster: Health and pandemic management; Green cluster: The disease and its pathophysiology; Blue cluster: Clinical epidemiology of the disease; Yellow cluster: Treatment of the disease.

Figure 2 (below): The research focus in a later stage of the Covid-19 literature

Figure 2 notes: Red cluster: Pandemic of viral pneumonia; public health and world spread; Green cluster: Treatment of COVID-19; Blue cluster: Epidemiology of the pandemic and disease; Yellow cluster: Surveys of disease and mental health; Purple cluster: Lab-work and diagnostics.

Tables of most central keywords in each cluster; in literature published during the early phase of the pandemic, and literature published during August.

The tables below provide an overview of the most central keywords in each cluster in the two time-periods. We extracted more keywords from the second time period as there was substantially more literature published in August alone, than the first 12 weeks of 2020, which the first map is based on. (Centrality measure used is weighted degree centrality)

Table 1: Most frequent keywords in the very early Covid-19 literature

Table 2 (below): Most frequent keywords in a later stage of the Covid-19 literature

To open the maps directly in VOSviewer, please use the relevant MAP and NET files (Co-occurrence files from March 23rd and the files for August2020Total, available through the Open Science Foundation (OSF) repository:
https://osf.io/54gqw/ 

Corpus downloaded from PubMed September 21st include 52105 publications. There were 3258 keywords that occurred in five or more manuscripts. The keywords divided into five main clusters. To create the map, we used the LinLog method. The minimum cluster size was set to 30.

Table of the most central keywords in each cluster

(Centrality measure used is weighted degree centrality/total link strength)

To open the map directly in VOSviewer, please use these the latest MAP and NET files, available through the Open Science Foundation (OSF) repository:
https://osf.io/54gqw/ 

In this week’s update, we are presenting something slightly different; the current focus of the COVID-19 literature, based on publications registered in PubMed as published during week 35-37 (August 24th-September 13th). This resulted in 8144 publications. There were 687 keywords that occurred in five or more manuscripts. These divided into four main clusters. To create the map, we used the LinLog method. The minimum cluster size was set to 15.

Table of the most central keywords in each cluster

(Centrality measure used is weighted degree centrality/total link strength)

To open the map directly in VOSviewer, please use the “KeywordCoOccurrance5Inst1Resol1_20200914FocusAug24thToSep13th” MAP and NET files, available through the Open Science Foundation (OSF) repository:
https://osf.io/54gqw/

Corpus downloaded from PubMed September 7th 48067 publications. There were 2960 keywords that occurred in five or more manuscripts. The keywords divided into five main clusters. To create the map, we used the LinLog method. The minimum cluster size was set to 30.

Table of the most central keywords in each cluster

(Centrality measure used is weighted degree centrality/total link strength)

To open the map directly in VOSviewer, please use these the latest MAP and NET files, available through the Open Science Foundation (OSF) repository:
https://osf.io/54gqw/ 

Corpus downloaded from PubMed August 31st 46037 publications. There were 2833 keywords that occurred in five or more manuscripts. The keywords divided into five main clusters. To create the map, we used the LinLog method. The minimum cluster size was set to 30.

Table of the most central keywords in each cluster

(Centrality measure used is weighted degree centrality/total link strength)

To open the map directly in VOSviewer, please use these the latest MAP and NET files, available through the Open Science Foundation (OSF) repository:
https://osf.io/54gqw/ 

Corpus downloaded from PubMed August 24th 41913 publications. There were 2516 keywords that occurred in five or more manuscripts. The keywords divided into five main clusters. To create the map, we used the LinLog method. The minimum cluster size was set to 30.

Table of the most central keywords in each cluster

(Centrality measure used is weighted degree centrality/total link strength)

To open the map directly in VOSviewer, please use these the latest MAP and NET files, available through the Open Science Foundation (OSF) repository:
https://osf.io/54gqw/ 

Corpus downloaded from PubMed August 17th 39914 publications. There were 2395 keywords that occurred in five or more manuscripts; this is the same number as last week. However, the order of the most frequently occurring keywords has changed slightly. The keywords divided into five main clusters. To create the map, we used the LinLog method. The minimum cluster size was set to 30.

Table of the most central keywords in each cluster

(Centrality measure used is weighted degree centrality/total link strength)

To open the map directly in VOSviewer, please use these the latest MAP and NET files, available through the Open Science Foundation (OSF) repository:
https://osf.io/54gqw/ 

Corpus downloaded from PubMed August 10th 38271 publications. There were 2395 keywords that occurred in five or more manuscripts. These divided into five main clusters. To create the map, we used the LinLog method. The minimum cluster size was set to 30.

Table of the most central keywords in each cluster

(Centrality measure used is weighted degree centrality/total link strength)

To open the map directly in VOSviewer, please use these the latest MAP and NET files, available through the Open Science Foundation (OSF) repository:
https://osf.io/54gqw/ 

Corpus downloaded from PubMed August 3rd 36201 publications. There were 2262 keywords that occurred in five or more manuscripts. These divided into five main clusters. To create the map, we used the LinLog method. The minimum cluster size was set to 25.

Table of the most central keywords in each cluster

(Centrality measure used is weighted degree centrality/total link strength)

To open the map directly in VOSviewer, please use these the latest MAP and NET files, available through the Open Science Foundation (OSF) repository:
https://osf.io/54gqw/ 

Corpus downloaded from PubMed July 27th: 33970 publications. There were 2111 keywords that occurred in five or more manuscripts. These divided into five main clusters. To create the map, we used the LinLog method.

Table of the most central keywords in each cluster

(Centrality measure used is weighted degree centrality/total link strength)

To open the map directly in VOSviewer, please use these the latest MAP and NET files, available through the Open Science Foundation (OSF) repository:
https://osf.io/54gqw/ 

In this week’s update, we are presenting something slightly different; the current focus of the COVID-19 literature, based on publications registered in PubMed as published in the past 3 weeks (21 days). This resulted in 6001 publications. There were 677 keywords that occurred in five or more manuscripts. These divided into five main clusters. To create the map, we used the LinLog method. The minimum cluster size was set to 15.

• Cluster 1 (red) covers literature focusing on Public health strategies, treatment, and the global impact of COVID-19.
• Cluster 2 (green) includes literature on the role of existing drugs and vaccines in the current pandemic.
• Cluster 3 (blue) focuses on demographics and risk factors.
• Cluster 4 (yellow) covers literature concentrated on understanding the nature of the virus.
• Cluster 5 (Purple) addresses the association between COVID-19 and mental health.

We find it interesting that there does not seem to be a big shift in publication focus from earlier in the pandemic; however, expect this to change as more time has elapsed, allowing for other types of studies to mature and be ready for publication, especially on topics relating to social consequences of the pandemic.

Table of the most central keywords in each cluster

(Centrality measure used is weighted degree centrality/total link strength)

To open the map directly in VOSviewer, please use the latest MAP and NET files, available through the Open Science Foundation (OSF) repository:
https://osf.io/54gqw/

Corpus downloaded from PubMed July 13th: 30116 publications. There were 1835 keywords that occurred in five or more manuscripts. These divided into five main clusters. To create the map, we used the LinLog method. The minimum cluster size was set to 25.

Table of the most central keywords in each cluster

(Centrality measure used is weighted degree centrality/total link strength)

Cluster description in progress

To open the map directly in VOSviewer, please use these the latest MAP and NET files, available through the Open Science Foundation (OSF) repository:
https://osf.io/54gqw/ 

Corpus downloaded from PubMed July 6th: 28256 publications. There were 1732 keywords that occurred in five or more manuscripts. These divided into five main clusters. To create the map, we used the LinLog method.

Table of the most central keywords in each cluster

(Centrality measure used is weighted degree centrality/total link strength)

Cluster description in progress

To open the map directly in VOSviewer, please use these the latest MAP and NET files, available through the Open Science Foundation (OSF) repository:
https://osf.io/54gqw/ 

Corpus downloaded from PubMed June 29th: 26187 publications. There were 1789 keywords that occurred in five or more manuscripts. These divided into five main clusters. To create the map, we used the LinLog method.

Table of the most central keywords in each cluster

(Centrality measure used is weighted degree centrality/total link strength)

Cluster description in progress

To open the map directly in VOSviewer, please use these the latest MAP and NET files, available through the Open Science Foundation (OSF) repository:
https://osf.io/54gqw/ 

Corpus downloaded from PubMed June 22nd: 24167 publications. There were 1656 keywords that occurred in five or more manuscripts. These divided into five main clusters. To create the map, we used the LinLog method.

Table of the most central keywords in each cluster

(Centrality measure used is weighted degree centrality/total link strength)

Cluster description in progress

To open the map directly in VOSviewer, please use these the latest MAP and NET files, available through the Open Science Foundation (OSF) repository:
https://osf.io/54gqw/ 

Corpus downloaded from PubMed June 15th: 22147 publications. There were 1535 keywords that occurred in five or more manuscripts. These divided into five main clusters. To create the map, we used the LinLog method.

Table of the most central keywords in each cluster

(Centrality measure used is weighted degree centrality/total link strength)

Cluster description in progress

To open the map directly in VOSviewer, please use these two files.
KeywordCoOccurrance5Inst1Resol1_20200615NETb
KeywordCoOccurrance5Inst1Resol1_20200615MAPb

The maps are created with the following “thesaurus” to clean the data:
Thesaurus20200601

Corpus downloaded from PubMed June 8th: 19426 publications. There were 1814 keywords that occurred in four or more manuscripts. These divided into five main clusters. To create the map, we used the LinLog method.

Table of the most central keywords in each cluster

(Centrality measure used is weighted degree centrality/total link strength)

Cluster description in progress

To open the map directly in VOSviewer, please use these two files.
KeywordCoOccurrance4Inst1Resol1_20200608NET
KeywordCoOccurrance4Inst1Resol1_20200608MAP

The maps are created with the following “thesaurus” to clean the data:
Thesaurus20200601

Corpus downloaded from PubMed 1st June: 17346 publications. There were 1648 keywords that occurred in four or more manuscripts. These divided into five main clusters.

Table of the most central keywords in each cluster

(Centrality measure used is weighted degree centrality/total link strength)

Cluster description in progress

To open the map directly in VOSviewer, please use these two files.
CoOccurrance4Inst1Resol1_20200601MAP
CoOccurrance4Inst1Resol1_20200601NET

The maps are created with the following “thesaurus” to clean the data:
Thesaurus20200601

Corpus downloaded from PubMed 25th May: 15438 publications. There were 1515 keywords that occurred in four or more manuscripts. These divided into five main clusters.

Table of the most central keywords in each cluster

(Centrality measure used is weighted degree centrality)

Cluster description in progress

To open the map directly in VOSviewer, please use these two files.
KeywordCoOccurrance4Inst1Resol20200525MAP
KeywordCoOccurrance4Inst1Resol20200525NET

The maps are created with the following “thesaurus” to clean the data:
KeywordCoOccurrance_20200427_Thesaurus

Corpus downloaded from PubMed 18th May: 13103 publications. There were 1861 keywords that occurred in three or more manuscripts. These divided into four main clusters.

This week’s rapport strengthens our observation from last couple of weeks that the clusters and topics seem to vary less, signifying that cluster may be more robust. This is true even if cluster 1 and 2 have changed places. Maybe this switch could imply that research and publication is switching more to how to managing the disease, than mainly be focused on the understanding of the phenomenon. In cluster 3 the term(s) (hydroxyl)chloroquine is  rising on the list. Last week there was a fifth cluster that we called “clinical issues”, but the terms included in this cluster have returned to the third cluster where they originated.

Cluster 1 could be called “disease epidemiology” and includes terms like pneumonia, epidemiology, outbreak, infection, infectious diseases, novel coronavirus pneumonia, treatment and transmission. This cluster contains 2 major topics: a) the main term “the disease” which in addition to the above mentioned terms, but also b) “inpatient care” which contains terms like ARDS, critical illness and airway management. In this cluster some smaller topics are also present.

Cluster 2 could be called «global virus pandemic» and includes terms (beta)coronavirus infection, viral pneumonia, pandemic/epidemic, public health, infection control and quarantine. This cluster contains 3 main topics a) “preventive epidemiology”, which in addition to the terms public health and quarantine contains terms like disease outbreak, travel and contact tracing, b) “care management” which includes terms like infection control, practice guidelines and health personnel and c) “vulnerability” that contains terms like mental health, psychological stress and health service accessibility.

Cluster 3 could be called “pathophysiology and origin” and includes terms like SARS, animals, zoonosis, antiviral agents, viral genome, phylogeny and chiroptera. This cluster contains to major topics: a) “viral biology” which in addition to terms like SARS, animals, viral genome and phylogeny includes terms like spike glycoprotein and peptidyl-dipeptidase a and b) “drug treatment” which in addition to the term antiviral agent, (hydroxyl)chloroquine and chiroptera contains terms like traditional Chinese medicine and lopinavir.

Cluster 4 could be named “clinical epidemiology” and contains terms like mortality, age, gender, clinical laboratory techniques and fever. Also in this cluster, there is only this major topic. Smaller topics include “population care”, “serology” and “pregnancy”.

Table of the most central keywords in each cluster

(Centrality measure used is weighted degree centrality)

Cluster description in progress

To open the map directly in VOSviewer, please use these two files.

KeywordCoOccurrance3Inst1Resol20200518MAP
KeywordCoOccurrance3Inst1Resol20200518NET

The maps are created with the following “thesaurus” to clean the data:
KeywordCoOccurrance_20200427_Thesaurus

Corpus downloaded from PubMed 11th May: 11005 publications. There were 1634 keywords that occurred in three or more manuscripts. These divided into four main clusters.

Table of the most central keywords in each cluster

(Centrality measure used is weighted degree centrality)

To open the map directly in VOSviewer, please use these two files.

KeywordCoOccurrance_20200511_NET
KeywordCoOccurrance_20200511_MAP

The maps are created with the following “thesaurus” to clean the data:
KeywordCoOccurrance_20200427_Thesaurus