Content and sentiment analysis of gabapentinoid-related tweets: An infodemiology study

How to apply zero-shot learning to text data in substance use research: An overview and tutorial with media data

A vast amount of media-related text data is generated daily in the form of social media posts, news stories or academic articles. These text data provide opportunities for researchers to analyse and understand how substance-related issues are being discussed. The main methods to analyse large text data (content analyses or specifically trained deep-learning models) require substantial manual annotation and resources. A machine-learning approach called 'zero-shot learning' may be quicker, more flexible and require fewer resources. Zero-shot learning uses models trained on large, unlabelled (or weakly labelled) data sets to classify previously unseen data into categories on which the model has not been specifically trained. This means that a pre-existing zero-shot learning model can be used to analyse media-related text data without the need for task-specific annotation or model training. This approach may be particularly important for analysing data that is time critical. This article describes the relatively new concept of zero-shot learning and how it can be applied to text data in substance use research, including a brief practical tutorial.

General practitioners' decision-making process to prescribe pain medicines for low back pain: a qualitative study

BACKGROUND

Pain medicines are widely prescribed by general practitioners (GPs) when managing people with low back pain (LBP), but little is known about what drives decisions to prescribe these medicines.

OBJECTIVES

The aim of this study was to investigate what influences GPs' decision to prescribe pain medicines for LBP.

DESIGN

Qualitative study with in-depth interviews.

SETTING

Australian primary care.

PARTICIPANTS

We interviewed 25 GPs practising in Australia experienced in managing LBP (mean (SD) age 53.4 (9.1) years, mean (SD) years of experience: 24.6 (9.3), 36% female). GPs were provided three vignettes describing common LBP presentations (acute exacerbation of chronic LBP, subacute sciatica and chronic LBP) and were asked to think aloud how they would manage the cases described in the vignettes.

DATA ANALYSIS

We summarised GP's choices of pain medicines for each vignette using content analysis and used framework analysis to investigate factors that affected GP's decision-making.

RESULTS

GPs more commonly prescribed opioid analgesics. Anticonvulsants and antidepressants were also commonly prescribed depending on the presentation described in the vignette. GP participants made decisions about what pain medicines to prescribe for LBP largely based on previous experiences, including their own personal experiences of LBP, rather than guidelines. The choice of pain medicine was influenced by a range of clinical factors, more commonly the patient's pathoanatomical diagnosis. While many adhered to principles of judicious use of pain medicines, polypharmacy scenarios were also common. Concerns about drug-seeking behaviour, adverse effects, stigma around opioid analgesics and pressure from regulators also shaped their decision-making process.

CONCLUSIONS

We identified several aspects of decision-making that help explain the current profile of pain medicines prescribed for LBP by GPs. Themes identified by our study could inform future implementation strategies to improve the quality use of medicines for LBP.