
==== Front
BMJ Open
BMJ Open
bmjopen
bmjopen
BMJ Open
2044-6055
BMJ Publishing Group BMA House, Tavistock Square, London, WC1H 9JR

10.1136/bmjopen-2024-084728
bmjopen-2024-084728
Protocol
Addiction
1681
1506
AI and Big Data approaches to addressing the opioid crisis: a scoping review protocol
Amjad Maaz 1maaz.amjad@austin.utexas.edu

https://twitter.com/SScottGraham_
http://orcid.org/0000-0003-1569-2428
Graham Scott 2ssg@utexas.edu

McCormick Katie 1kmccormick@utexas.edu

https://twitter.com/KaseyClabornPhD
http://orcid.org/0000-0002-8415-3869
Claborn Kasey 1kasey.claborn@austin.utexas.edu

1 Steve Hicks School of Social Work, The University of Texas at Austin, Austin, Texas, USA
2 Rhetoric & Writing, The University of Texas at Austin, Austin, Texas, USA
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.

KC has been retained by the plaintiff in the opioid litigation to serve as an expert advisor on matters pertaining to health information and technology. All other authors declare that they have no conflicts of interest.

Dr; kasey.claborn@austin.utexas.edu
2024
31 8 2024
14 8 e08472826 1 2024
13 8 2024
Copyright © Author(s) (or their employer(s)) 2024. Re-use permitted under CC BY-NC. No commercial re-use. See rights and permissions. Published by BMJ.
2024
https://creativecommons.org/licenses/by-nc/4.0/ This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.

Abstract

Introduction

This paper outlines the steps necessary to assess the latest developments in artificial intelligence (AI) as well as Big Data technologies and their relevance to the opioid crisis. Fatal opioid overdoses have risen to over 82 998 annually in the USA. This highlights the need for urgent and effective data-driven solutions. AI approaches, such as machine learning, deep learning and natural language processing, have been employed to analyse patterns and trends in overdose data and facilitate timely interventions. However, a comprehensive scoping review on the effectiveness of AI-driven technologies to detect, treat, prevent or respond to the opioid crisis remains absent. Thus, it is important to identify recent advancements in AI and Big Data technologies in addressing the opioid crisis.

Methods and analysis

We will electronically search four scientific databases (PubMed, Web of Science, Engineering Village and PsycInfo), including finding reference lists and grey literature from 2013 to 2023. Covidence will be used for screening and selecting papers. We will extract information such as citation details, study context, data used, AI/Big Data technologies, features, algorithms and evaluation metrics. This data will be synthesised, analysed and summarised to draw meaningful conclusions and identify future directions to tackle the opioid crisis.

Ethics and dissemination

Ethics approval is not required. Results will be disseminated via conference presentations and peer-reviewed publication.

Natural Language Processing
Substance misuse
Machine Learning
Health informatics
Artificial Intelligence
==== Body
pmcStrengths and limitations of this study

The review will synthesise literature on the surveillance, detection, treatment, prevention and long-term recovery of opioid use disorder and overdose in North America.

The review explores artificial intelligence-based interventions reported in both peer-reviewed and grey literature to address the opioid crisis in North America.

The scoping review excludes literature published in languages other than English.

The findings of the review are specific to North America and may not be generalisable to other regions.

Introduction

Drug overdose deaths in the USA have skyrocketed over the past two decades, creating an unprecedented public health crisis that has devastated communities across the nation. Between June 2022 and June 2023, over 106 500 individuals died from a drug overdose in the USA.1 Opioids are the primary driver of overdose deaths, accounting for more than 80 000 of those fatalities2; and these numbers are expected to continue to rise with the increasing prevalence of synthetic opioids (eg, fentanyl). Compounding the staggering number of opioid-involved overdose deaths is the countless number of individuals suffering from opioid use disorder (OUD). Though evidence-based medication-assisted treatments exist (eg, methadone, buprenorphine, naltrexone),3 4 a large number of individuals remain undiagnosed, and of those who are diagnosed, only 10% receive treatment.5 This lack of diagnosis and treatment has contributed to a significant rise in opioid overdose deaths,5 6 as well as over-utilisation of emergency departments for those seeking care. In fact, opioid-related admissions to emergency departments have soared, with some estimates exceeding 2.8 million in a single year,7 which has only increased with the onset of the COVID-19 pandemic.8

The alarming impact of the opioid and overdose crisis highlights an urgent need for efficient, cost-effective solutions to tackle the overdose crisis. Several strategies, such as predictive modelling for identifying individuals at risk of developing OUD,9 surveillance systems to track the spread and analyse patterns of opioid use and overdose incidents,10 healthcare systems for optimised patient engagement and personalised treatment plans and improving long-term recovery outcomes,11 are being explored to efficiently address the opioid crisis.12 Hence, it is imperative to identify and analyse the existing literature on artificial intelligence (AI) and Big Data technologies/interventions to address the opioid crisis, particularly highlighting the existing solutions based on machine learning (ML), natural language processing (NLP) and related informatics technologies that have the potential to mitigate the effects of this public health emergency.

In recent years, the application of AI in healthcare and medicine has surged, with ML playing a pivotal role in public health interventions.1317 Given the pressing need to address the opioid crisis, many studies have harnessed AI and informatics technologies as potential solutions towards monitoring drug overdose incidents,18 identifying emerging trends in overdose fatalities,19 predicting overdose risks,20 identifying OUD11 21 and creating demographic profiles of individuals who have died due to opioid overdose.22 These studies exemplify the potential of AI and Big Data to make important contributions towards alleviating the opioid crisis.

As AI and informatics technologies are deployed to address the opioid crisis, researchers are presented with a rapidly evolving and dynamic landscape of research and innovation. Moreover, numerous interventions are being developed by technology companies that may not submit their work for peer-reviewed publication in the biomedical and public health literature. Consequently, a synthesis of existing efforts using AI and Big Data tools for OUD and overdose can provide a critical roadmap for guiding future research. Thus, the primary objective of the review is to identify and describe advancements in AI and Big Data technologies to address OUD and opioid overdose. By doing so, we aim to highlight current research gaps and offer valuable insights to inform future efforts in early detection, treatment and prevention of OUD, while also guiding future endeavours in overdose prevention and response.

Why it is important to do this review

Consolidating existing literature is an essential first step in fully understanding and comprehending the application and potential of AI and Big Data tools in addressing the opioid crisis. A scoping review is particularly beneficial in this context, as it provides a consolidated overview of the current initiatives in this area. By conducting a thorough scoping review and narrative evidence synthesis, we can effectively shape future endeavours. This can be achieved by documenting the range of existing initiatives, elucidating the computational and methodological foundations of the work and identifying key gaps in current research efforts.

Review objectives

The primary objective of the scoping review is to examine how AI and Big Data technologies have been applied to tackle the opioid crisis, with specific goals to:

Analyse peer-reviewed and grey literature on the opioid crisis to identify advancements in AI and Big Data technologies for responding to and preventing opioid overdose events.

Examine and analyse the datasets, features, types of ML algorithms and their evaluation metrics presented in scientific manuscripts pertaining to the opioid crisis.

Review and analyse the existing literature on AI and Big Data technologies/interventions used to address the issues related to the opioid crisis, with an aim to highlight current detection, treatment and intervention solutions for both opioid-involved overdose and OUD. This will pave the way for future research in this area, with the potential to save lives and mitigate this public health crisis.

Methods and analysis

This review outlines the protocol for an ongoing scoping review, conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses Scoping Review Extension (PRISMA-ScR) guidelines.23 A completed checklist will be made available on study completion to ensure rigour and facilitate replicability. By adhering to these established guidelines,23 the scoping review aims to maintain a high level of methodological quality while exploring existing studies that propose AI and Big Data interventions for the opioid crisis. The key steps of the scoping review include: (a) developing and conducting an appropriate search strategy; (b) screening results to select relevant studies; (c) extracting pertinent data; (d) synthesising the data; and (e) summarising and presenting the findings. Through a systematic and thorough approach, the scoping review aims to provide valuable insights that can inform future initiatives in AI for research, policy-making and clinical practice to address the opioid crisis. We have registered this protocol with the Open Science Framework.1

Search strategy for identification of studies and grey literature

The search strategy was developed in collaboration with a university behavioural science librarian. The third wave of the opioid crisis in North America began in 2013, and the USA started experiencing a dramatic increase in opioid-related overdose deaths primarily due to synthetic opioids (eg, fentanyl); therefore, we confined the search and only included studies conducted from 2013 onwards. To find relevant peer-reviewed scientific literature, four databases will be used to search for papers published between 2013 and 2023: (a) PubMed, (b) Web of Science, (c) Engineering Village and (d) PsycInfo. These databases were selected because they cover a wide range of relevant literature from various disciplines, including medicine, engineering, psychology and general sciences. The search will be conducted in English, employing the following search terms: (“opioid” OR “opioid use disorder” OR “OUD” OR “opioid addiction” OR “overdose”) AND (“artificial intelligence” OR “AI” OR “A.I.” OR “machine learning” OR “deep learning” OR “NLP” OR “Natural Language Processing”). The specific search terms for each database are described in online supplemental file 1.

The inclusion of grey literature is essential to gaining a comprehensive understanding of available evidence regarding the use of AI and Big Data technologies for the opioid crisis.24 The goal of this search is to locate pertinent websites, databases and documents through a two-step process. In the first step, we will search Google for potential websites, from which we will compile a list that may contain pertinent documents. Initially, we will conduct searches on Google to identify relevant organisations using keywords such as (“artificial intelligence” OR “Big Data” OR “machine learning”) AND (“opioid interventions” OR “opioid crisis” AND “public health” AND “substance abuse” OR “substance abuse prevention”). In the second step, two reviewers will manually perform a focused search within these websites to identify records that fulfil the inclusion criteria.

Furthermore, considering the rapid innovations in AI and Big Data we will use Google Patents25 to identify relevant patents or registered intellectual property. Additionally, we will manually search the reference lists of studies that meet the inclusion criteria, using well-known sources like Government Websites (CDC, NIH, NIDA, National Library of Medicine), Healthcare Innovation,26 Health Care IT news,27 Medium,28 American Public Health Association,29 MedlinePlus,30 Alcohol and Drug Foundation,31 Mayo Clinic Press32 and other reputable media outlets reporting on AI products. This comprehensive approach will ensure that we gather exhaustive and diverse grey literature sources, including government, non-government organisations and private sector innovations. This is crucial to creating a complete picture of available evidence and insights.33

Patient and public involvement

This research was designed, conducted, reported and planned for dissemination without the involvement of patients or the general public.

Inclusion/exclusion criteria

To pinpoint pertinent studies, we have established key inclusion and exclusion criteria in accordance with the population, concepts and context recommended by PRISMA-ScR guidelines (see online supplemental file 2).23 We will use Covidence,34 a software platform for managing systematic reviews, to import all articles found through database searches. This platform facilitates the handling of article deduplication, abstract assessment and full-text review. In the initial screening, titles and abstracts of all the articles will be reviewed against the inclusion/exclusion criteria by two trained research assistants and supervised by a doctoral-level researcher. In case of discrepancies, a third reviewer (doctoral-level researcher) will review those articles to resolve discrepancies. Once the initial screening is done, two reviewers will perform a full-text review for detailed information extraction from studies that meet the initial screening criteria. During the full-text review, the two reviewers will carefully examine the entire texts of initially selected studies that meet the screening criteria using Covidence.34 This approach ensures a systematic and focused identification of studies pertinent to the objectives and scope of the review. By applying these criteria, the review aims to compile a diverse collection of studies that elucidate the application of AI and Big Data technologies in addressing the opioid crisis. The specific inclusion and exclusion criteria for this review are listed in table 1.

Table 1 Review eligibility criteria

	Inclusion criteria	Exclusion criteria	
Population	Studies that include human subjects (of any age, race, or gender) who have experienced an overdose or have a past or present diagnosis of OUD or are at-risk for OUD or overdose and use applications of AI and Big Data technology-based interventions	Studies that focus on animal models or exclude human subjects, and do not discuss applications of AI and Big Data technology-based interventions for OUD (not human population)	
Concepts/Topic	Studies that refer to the concepts and applications of AI and Big Data technologies to address the opioid crisis	Studies that do not discuss AI and Big Data technologies in the opioid crisis context (focus misalignment)	
Source	Full-text articles

Full-text conference proceedings

Studies that develop or evaluate the use of AI/ML systems to address the opioid crisis

Studies that focus on the use of AI and Big Data technologies/interventions to address the opioid crisis

Studies that are only focused on the opioid crisis or studies where the opioid-related data could be disaggregated

Grey literature sources (Google Scholar, Google Patents) to identify potentially eligible articles that discuss the use of AI and Big Data technologies/interventions for the opioid crisis

	Articles that are not empirical and do not use AI/ML systems to address the opioid crisis (not empirical)

Studies irrelevant to AI and Big Data technologies (technology irrelevance)

Dissertations, books, opinion papers, abstracts, posters and preprints

Articles lacking full-text availability or with a misalignment of focus (eg, use genome data, or focus on topics unrelated to the opioid crisis or AI/Big Data technologies) (focus misalignment)

Studies that do not use data from the USA or Canada

	
Language	Articles written in English	Articles that are not available in the English language	
Study limits	Studies published between 1 January 2013 and 12 May 2023	Studies published before 1 January 2013 or after 12 May 2023	
AIartificial intelligenceMLmachine learningOUDopioid use disorder

Data extraction and management

We will create two separate spreadsheets (eg, MS Excel): one for extracting information from the peer-reviewed studies and one for extracting information from the grey literature. A doctoral-level researcher and two trained research assistants will perform data extraction. The doctoral-level researcher and one trained research assistant will extract information from each included study, comparing the information to ensure that all required data meeting the screening criteria align with the objectives and scope of the review. A similar approach will be followed for the grey literature.

The data from studies included in the review (selected after full-text review) will be recorded in a data extraction spreadsheet. In the peer-reviewed spreadsheet, we will include categories such as citation information (eg, authors, title, publication date), study context and setting (eg, the country, target population and intervention role), the data used in the studies (eg, medical records, social media data or prescription data) and the AI/Big Data technology employed (eg, AI/ML tools, Big Data sources and software). It will also detail the specific features used (eg, text features, numerical features or multimodal data), types of algorithms (eg, ML, deep learning (DL) or neural networks), and evaluation metrics (eg, precision, recall, accuracy and F1 score). We will review the conflict of interest section of the included manuscripts, and if the manuscripts report patents or express intent to submit patent applications, we will also extract this information. For the grey literature spreadsheet, we will include categories such as product company (eg, product owner or developer), technology used (eg, AI/Big Data) and product functionality (eg, monitoring, treatment, prediction of high-risk OUD patients). Table 2 shows a detailed breakdown of the categories and fields used in data extraction.

Table 2 Data extraction categories and fields

Publication details	Title, publication venue, authors, publication date, document type (research article, patent, grey literature)	
Study context and setting	Country, settings (hospital, community, online), target population (patients with opioid use disorder, healthcare providers, general public)	
AI/Big Data technology	AI/ML tools (ML, predictive modelling, NLP), data sources (EHRs, social media), data split for modelling, Big Data tools/software, feature engineering, ground truth approach, algorithms, evaluation metrics	
Results, applications and health outcomes	Narrative review of findings and impact of technology on health outcomes	
AIartificial intelligenceEHRselectronic health recordsMLmachine learningNLPnatural language processing

Data synthesis

The scoping review aims to provide a comprehensive understanding of how AI and Big Data technologies have been suggested as tools for addressing the opioid crisis. Through a narrative lens, the data synthesis derived from the assembled scholarly articles highlight emerging trends, uncover knowledge gaps and identify commonalities in how AI and Big Data technologies have been deployed for the opioid crisis. This includes their use in detecting fatal and non-fatal overdose incidents, diagnosing OUD, aiding in treatment and supporting prevention and recovery efforts. We will thoroughly synthesise and summarise information regarding the AI and Big Data techniques featured in existing studies to address the opioid crisis. At the end of the study, we will provide a written summary of the data extracted from the reviewed literature.

Expected outcomes

The expected outcome of the scoping review is a comprehensive analysis of how AI and Big Data technologies have been deployed to address the opioid crisis. The review will catalogue the types of AI and Big Data tools outlined in the scientific literature related to the opioid crisis and perform a thorough analysis of the data sets, features, ML algorithms and evaluation metrics encompassed in these studies. Furthermore, the scoping review seeks to scrutinise existing body of literature to illuminate the detection, treatment and prevention along with other intervention strategies in practice and provide direction for future research in confronting the opioid crisis. The outcome of the scoping review will focus on highlighting recent advancements in AI and Big Data technologies and their tangible applications for detecting fatal and non-fatal overdose incidents and OUD, personalising treatment for OUD and reducing overdose fatalities. The selected studies will be synthesised, analysed and summarised to draw meaningful conclusions and recommendations to address the urgent challenge of the opioid crisis. These insights will provide a comprehensive review of existing interventions using AI and Big Data technologies to inform further development and utilisation of such interventions to address the worsening opioid crisis.

Ethics and dissemination

This review consolidates knowledge from various published sources. As the data used for this review is not associated with specific individuals or patients, no personal safety is jeopardised, negating the need for formal ethical approval. Additionally, the conclusions drawn from this review will be disseminated via peer-reviewed journals and academic conference presentations.

supplementary material

10.1136/bmjopen-2024-084728 online supplemental file 1

10.1136/bmjopen-2024-084728 online supplemental file 2

Review Process File
31 08 2024

https://doi.org/10.17605/OSF.IO/3V7P6.

Funding: The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

Prepublication history and additional supplemental material for this paper are available online. To view these files, please visit the journal online (https://doi.org/10.1136/bmjopen-2024-084728).

Provenance and peer review: Not commissioned; externally peer reviewed.

Patient consent for publication: Not applicable.

Patient and public involvement: Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.
==== Refs
References

1 Ahmad FB Cisewski JA Rossen LM et al Provisional drug overdose death counts 2023 Available https://www.cdc.gov/nchs/nvss/vsrr/drug-overdosedata.htm Accessed 17-Nov-2023
2 Centers for Disease Control and Prevention Provisional data shows U.S. drug overdose deaths top 100,000 in 2022 18 05 2023 Available https://blogs.cdc.gov/nchs/2023/05/18/7365/ Accessed 9-Jul-2023
3 Connery HS Medication-assisted treatment of opioid use disorder: review of the evidence and future directions Harv Rev Psychiatry 2015 23 63 75 10.1097/HRP.0000000000000075 25747920
4 Kampman K Jarvis M American Society of Addiction Medicine (ASAM) national practice guideline for the use of medications in the treatment of addiction involving opioid use J Addict Med 2015 9 358 67 ADM.0000000000000166 10.1097/ADM.0000000000000166 26406300
5 Campo DS Gussler JW Sue A et al Accurate spatiotemporal mapping of drug overdose deaths by machine learning of drug-related web-searches PLoS One 2020 15 e0243622 10.1371/journal.pone.0243622 33284864
6 Jalal H Buchanich JM Roberts MS et al Changing dynamics of the drug overdose epidemic in the United States from 1979 through 2016 Science 2018 361 eaau1184 10.1126/science.aau1184 30237320
7 Langabeer JR Stotts AL Bobrow BJ et al Prevalence and charges of opioid-related visits to U.S. emergency departments Drug Alcohol Depend 2021 221 108568 10.1016/j.drugalcdep.2021.108568 33578297
8 Soares WE III Melnick ER Nath B et al Emergency department visits for nonfatal opioid overdose during the COVID-19 Pandemic Across Six US Health Care Systems Ann Emerg Med 2022 79 158 67 10.1016/j.annemergmed.2021.03.013 34119326
9 Bharat C Hickman M Barbieri S et al Big data and predictive modelling for the opioid crisis: existing research and future potential Lancet Digit Health 2021 3 e397 407 10.1016/S2589-7500(21)00058-3 34045004
10 Burkom H Loschen W Wojcik R et al Electronic Surveillance System for the Early Notification of Community-Based Epidemics (ESSENCE): overview, components, and public health applications JMIR Public Health Surveill 2021 7 e26303 10.2196/26303 34152271
11 Sokol R Tammaro E Kim JY et al Linking MATTERS: barriers and facilitators to implementing emergency department-initiated buprenorphine-naloxone in patients with opioid use disorder and linkage to long-term care Substance Use & Misuse 2021 56 1045 53 10.1080/10826084.2021.1906280 33825669
12 Neill DB Herlands W Machine learning for drug overdose surveillance J Technol Hum Serv 2018 36 8 14 10.1080/15228835.2017.1416511
13 Topol EJ High-performance medicine: the convergence of human and artificial intelligence Nat Med 2019 25 44 56 10.1038/s41591-018-0300-7 30617339
14 Beam AL Kohane IS Translating artificial intelligence into clinical care JAMA 2016 316 2368 9 10.1001/jama.2016.17217 27898974
15 Fix J Ising AI Proescholdbell SK et al Linking emergency medical services and emergency department data to improve overdose surveillance in North Carolina Public Health Rep 2021 136 54S 61S 10.1177/00333549211012400 34726971
16 Bozorgi P Porter DE Eberth JM et al The leading neighborhood-level predictors of drug overdose: A mixed machine learning and spatial approach Drug Alcohol Depend 2021 229 109143 10.1016/j.drugalcdep.2021.109143 34794060
17 Volkow ND Chandler RK Villani J Need for comprehensive and timely data to address the opioid overdose epidemic without a blindfold Addiction 2022 117 2132 4 10.1111/add.15957 35611646
18 Fisher A Young MM Payer D et al Automating detection of drug-related harms on social media: machine learning framework J Med Internet Res 2023 25 e43630 10.2196/43630 37725410
19 Ward PJ Young AM Slavova S et al Deep neural networks for fine-grained surveillance of overdose mortality Am J Epidemiol 2023 192 257 66 10.1093/aje/kwac180 36222700
20 Lo-Ciganic W-H Huang JL Zhang HH et al Evaluation of machine-learning algorithms for predicting opioid overdose risk among medicare beneficiaries with opioid prescriptions JAMA Netw Open 2019 2 e190968 10.1001/jamanetworkopen.2019.0968 30901048
21 Goldfine C Lai JT Lucey E et al Wearable and wireless mhealth technologies for substance use disorder Curr Addict Rep 2020 7 291 300 10.1007/s40429-020-00318-8 33738178
22 Palis H Bélair M-A Hu K et al Overdose deaths and the COVID-19 pandemic in British Columbia, Canada Drug Alcohol Rev 2022 41 912 7 10.1111/dar.13424 34908203
23 Tricco AC Lillie E Zarin W et al PRISMA Extension for Scoping Reviews (PRISMA-ScR): checklist and explanation Ann Intern Med 2018 169 467 73 10.7326/M18-0850 30178033
24 Zeng D Cao Z Neill DB et al Artificial intelligence–enabled public health surveillance—from local detection to global epidemic monitoring and control Artificial Intelligence in Medicine Cambridge, MA Academic Press 2021 437 53
25 Google Patents 2023 Available https://patents.google.com/ Accessed 9-Jul-2023
26 Healthcare Innovative 2023 Available https://www.hcinnovationgroup.com/about-us Accessed 9-Jul-2023
27 Health Care News 2023 Available https://www.healthcareitnews.com/ Accessed 9-Jul-2023
28 Medium 2023 Available https://medium.com/ Accessed 9-Jul-2023
29 American Public Health Association 2023 Available https://www.apha.org/ Accessed 9-Jul-2023
30 MedlinePlus 2023 Available https://medlineplus.gov/ Accessed 9-Jul-2023
31 Alcohol and Drug Foundation 2023 Available https://adf.org.au/ Accessed 9-Jul-2023
32 Mayo Clinic Press 2023 Available https://mcpress.mayoclinic.org/ Accessed 9-Jul-2023
33 Paez A Gray literature: an important resource in systematic reviews J Evid Based Med 2017 10 233 40 10.1111/jebm.12266 28857505
34 Covidence Systematic Review Software Melbourne, AU Veritas Health Innovation 2023 Available https://www.covidence.org/ accessed 9-May-2023
