Dataset Information

Using machine learning to investigate self-medication purchasing in England via high street retailer loyalty card data.

ABSTRACT: The availability alongside growing awareness of medicine has led to increased self-treatment of minor ailments. Self-medication is where one 'self' diagnoses and prescribes over the counter medicines for treatment. The self-care movement has important policy implications, perceived to relieve the National Health Service (NHS) burden, increasing patient subsistence and freeing resources for more serious ailments. However, there has been little research exploring how self-medication behaviours vary between population groups due to a lack of available data. The aim of our study is to evaluate how high street retailer loyalty card data can help inform our understanding of how individuals self-medicate in England. Transaction level loyalty card data was acquired from a national high street retailer for England for 2012-2014. We calculated the proportion of loyalty card customers (n ~ 10 million) within Lower Super Output Areas who purchased the following medicines: 'coughs and colds', 'Hayfever', 'pain relief' and 'sun preps'. Machine learning was used to explore how 50 sociodemographic and health accessibility features were associated towards explaining purchasing of each product group. Random Forests are used as a baseline and Gradient Boosting as our final model. Our results showed that pain relief was the most common medicine purchased. There was little difference in purchasing behaviours by sex other than for sun preps. The gradient boosting models demonstrated that socioeconomic status of areas, as well as air pollution, were important predictors of each medicine. Our study adds to the self-medication literature through demonstrating the usefulness of loyalty card records for producing insights about how self-medication varies at the national level. Big data offer novel insights that add to and address issues that traditional studies are unable to consider. New forms of data through data linkage may offer opportunities to improve current public health decision making surrounding at risk population groups within self-medication behaviours.

SUBMITTER: Davies A

PROVIDER: S-EPMC6242371 | biostudies-literature | 2018

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Using machine learning to investigate self-medication purchasing in England via high street retailer loyalty card data.

Davies Alec A Green Mark A MA Singleton Alex D AD

PloS one 20181119 11

The availability alongside growing awareness of medicine has led to increased self-treatment of minor ailments. Self-medication is where one 'self' diagnoses and prescribes over the counter medicines for treatment. The self-care movement has important policy implications, perceived to relieve the National Health Service (NHS) burden, increasing patient subsistence and freeing resources for more serious ailments. However, there has been little research exploring how self-medication behaviours var ...[more]

PMID: 30452481

Similar Datasets

Project description:This study aimed to describe street food and takeaway food purchasing patterns in Sarajevo and Banja Luka, throughout the day and by city location. A cross-sectional evaluation of street food and takeaway food customers was conducted in 2017. All eligible vending sites (n = 348) in the vicinity of selected markets and bus stops were included. Data on the food items purchased, and time and geographic location of the purchases were collected. A total of 755 customers purchased 929 food items. Takeaway venues showed higher customer influx (5.0 vs. 2.0 customers observed per 10 min of observation, p < 0.001) and buying rates (6.7 vs. 2.0 items bought per 10 min of observation, p < 0.001; 1.5 vs. 1.0 items bought per customer, p < 0.001) than street food sites. These rates were higher in city peripheries for street food venues, and in city centres for takeaway establishments. The purchase of industrial food products prevailed throughout the day in street food venues, whereas most takeaway purchases comprised homemade foods, with or without industrial beverages. The proportion of customers buying foods and beverages together was higher in takeaway venues (15.3% vs. 6.0%, p < 0.001), especially during lunchtime and in city centres. In street food vending sites, sweet and savoury snacks seemed to be preferred in the afternoon, whereas in takeaway food establishments, savoury pastries and main dishes were mostly purchased at breakfast or lunch, and bread during the morning. Soft drinks and industrial juices were frequently purchased in both types of vending site and at all hours of the day, particularly in the afternoon. Our findings provide an overview of street food and takeaway food buying habits and consumer demands in these cities, reflecting local food culture and dietary behaviours. The identification of the meal contexts and city sub-regions in which specific purchasing practices emerge point to potential priority targets. These insights can be useful when designing interventions adapted to the specificities of these food environments and the food habits of customers.

Project description:Current approaches to understanding medication ordering errors rely on relatively small manually captured error samples. These approaches are resource-intensive, do not scale for computerized provider order entry (CPOE) systems, and are likely to miss important risk factors associated with medication ordering errors. Previously, we described a dataset of CPOE-based medication voiding accompanied by univariable and multivariable regression analyses. However, these traditional techniques require expert guidance and may perform poorly compared to newer approaches. In this paper, we update that analysis using machine learning (ML) models to predict erroneous medication orders and identify its contributing factors. We retrieved patient demographics (race/ethnicity, sex, age), clinician characteristics, type of medication order (inpatient, prescription, home medication by history), and order content. We compared logistic regression, random forest, boosted decision trees, and artificial neural network models. Model performance was evaluated using area under the receiver operating characteristic curve (AUROC) and the area under the precision-recall curve (AUPRC). The dataset included 5,804,192 medication orders, of which 28,695 (0.5%) were voided. ML correctly classified voids at reasonable accuracy; with a positive predictive value of 10%, ~20% of errors were included. Gradient boosted decision trees achieved the highest AUROC (0.7968) and AUPRC (0.0647) among all models. Logistic regression had the poorest performance. Models identified predictive factors with high face validity (e.g., student orders), and a decision tree revealed interacting contexts with high rates of errors not identified by previous regression models. Prediction models using order-entry information offers promise for error surveillance, patient safety improvements, and targeted clinical review. The improved performance of models with complex interactions points to the importance of contextual medication ordering information for understanding contributors to medication errors.

Project description:BackgroundTo date, the evaluation of diet has mostly been based on questionnaires and diaries that have their limitations in terms of being time and resource intensive, and a tendency toward social desirability. Loyalty card data obtained in retailing provides timely and objective information on diet-related behaviors. In Finland, the market is highly concentrated, which provides a unique opportunity to investigate diet through grocery purchases.ObjectiveThe aims of this study were as follows: (1) to investigate and quantify the selection bias in large-scale (n=47,066) loyalty card (LoCard) data and correct the bias by developing weighting schemes and (2) to investigate how the degree of loyalty relates to food purchases.MethodsMembers of a loyalty card program from a large retailer in Finland were contacted via email and invited to take part in the study, which involved consenting to the release of their grocery purchase data for research purposes. Participants' sociodemographic background was obtained through a web-based questionnaire and was compared to that of the general Finnish adult population obtained via Statistics Finland. To match the distributions of sociodemographic variables, poststratification weights were constructed by using the raking method. The degree of loyalty was self-estimated on a 5-point rating scale.ResultsOn comparing our study sample with the general Finnish adult population, in our sample, there were more women (65.25%, 30,696/47,045 vs 51.12%, 2,273,139/4,446,869), individuals with higher education (56.91%, 20,684/36,348 vs 32.21%, 1,432,276/4,446,869), and employed individuals (60.53%, 22,086/36,487 vs 52.35%, 2,327,730/4,446,869). Additionally, in our sample, there was underrepresentation of individuals aged under 30 years (14.44%, 6,791/47,045 vs 18.04%, 802,295/4,446,869) and over 70 years (7.94%, 3,735/47,045 vs 18.20%, 809,317/4,446,869), as well as retired individuals (23.51%, 8,578/36,487 vs 31.82%, 1,414,785/4,446,869). Food purchases differed by the degree of loyalty, with higher shares of vegetable, red meat & processed meat, and fat spread purchases in the higher loyalty groups.ConclusionsIndividuals who consented to the use of their loyalty card data for research purposes tended to diverge from the general Finnish adult population. However, the high volume of data enabled the inclusion of sociodemographically diverse subgroups and successful correction of the differences found in the distributions of sociodemographic variables. In addition, it seems that food purchases differ according to the degree of loyalty, which should be taken into account when researching loyalty card data. Despite the limitations, loyalty card data provide a cost-effective approach to reach large groups of people, including hard-to-reach population subgroups.

Project description:BackgroundLoyalty card data automatically collected by retailers provide an excellent source for evaluating health-related purchase behavior of customers. The data comprise information on every grocery purchase, including expenditures on product groups and the time of purchase for each customer. Such data where customers have an expenditure value for every product group for each time can be formulated as 3D tensorial data.ObjectiveThis study aimed to use the modern tensorial principal component analysis (PCA) method to uncover the characteristics of health-related purchase patterns from loyalty card data. Another aim was to identify card holders with distinct purchase patterns. We also considered the interpretation, advantages, and challenges of tensorial PCA compared with standard PCA.MethodsLoyalty card program members from the largest retailer in Finland were invited to participate in this study. Our LoCard data consist of the purchases of 7251 card holders who consented to the use of their data from the year 2016. The purchases were reclassified into 55 product groups and aggregated across 52 weeks. The data were then analyzed using tensorial PCA, allowing us to effectively reduce the time and product group-wise dimensions simultaneously. The augmentation method was used for selecting the suitable number of principal components for the analysis.ResultsUsing tensorial PCA, we were able to systematically search for typical food purchasing patterns across time and product groups as well as detect different purchasing behaviors across groups of card holders. For example, we identified customers who purchased large amounts of meat products and separated them further into groups based on time profiles, that is, customers whose purchases of meat remained stable, increased, or decreased throughout the year or varied between seasons of the year.ConclusionsUsing tensorial PCA, we can effectively examine customers' purchasing behavior in more detail than with traditional methods because it can handle time and product group dimensions simultaneously. When interpreting the results, both time and product dimensions must be considered. In further analyses, these time and product groups can be directly associated with additional consumer characteristics such as socioeconomic and demographic predictors of dietary patterns. In addition, they can be linked to external factors that impact grocery purchases such as inflation and unexpected pandemics. This enables us to identify what types of people have specific purchasing patterns, which can help in the development of ways in which consumers can be steered toward making healthier food choices.

Project description:ObjectivesChanges in public sector service spending may influence food consumption. We make use of changing local authority (LA) expenditure in England to assess impacts on food purchasing. We examine total LA service spending and explore two potential pathways: highways and transport spending which may affect access to food; and housing service expenditure which may affect household resources available to purchase foods.DesignLongitudinal panel survey at the LA level (2008-2015) using fixed effects linear regression.Setting324 LAs in England.Main exposureExpenditure per capita on total LA services, highways and transport services, and housing services.Main outcome measuresLA area estimates of purchasing of fresh fruits and vegetables, high in fat, sugar and salt (HFSS) foods, and takeaways at home, expressed as a percentage of total food and drink expenditure.ResultsTotal LA service spending decreased by 17% on average between 2008 and 2015. A 10% decrease in total LA spending was associated with a 0.071 percentage point decrease in HFSS (95% CI -0.093 to -0.050) and a 0.015 percentage point increase in takeaways (95% CI 0.006 to 0.024). A 10% decrease in highways and transport expenditure was associated with a 0.006 percentage point decrease in fruit and vegetable purchasing (95% CI -0.009 to -0.002) and a 0.006 percentage point increase in takeaway purchasing (95% CI 0.001 to 0.010). These associations were seen in urban areas only when analyses were stratified by rural/urban area status. A 0.006 percentage point decrease in HFSS purchasing was also seen with a 10% decrease in housing expenditure (95% CI -0.010 to -0.002).ConclusionChanges in LA spending may have impacts on food purchasing which are evident at the area level. This suggests that in addition to more prominent impacts such as foodbank use, austerity measures may have mixed impacts on food purchasing behaviours among the wider population. Individual-level research is needed to further elucidate these relationships.

Project description:BackgroundOver-the-counter (OTC) medications are frequently used to self-care for nonspecific ovarian cancer symptoms prior to diagnosis. Monitoring such purchases may provide an opportunity for earlier diagnosis.ObjectiveThe aim of the Cancer Loyalty Card Study (CLOCS) was to investigate purchases of OTC pain and indigestion medications prior to ovarian cancer diagnosis in women with and without ovarian cancer in the United Kingdom using loyalty card data.MethodsAn observational case-control study was performed comparing purchases of OTC pain and indigestion medications prior to diagnosis in women with (n=153) and without (n=120) ovarian cancer using loyalty card data from two UK-based high street retailers. Monthly purchases of pain and indigestion medications for cases and controls were compared using the Fisher exact test, conditional logistic regression, and receiver operating characteristic (ROC) curve analysis.ResultsPain and indigestion medication purchases were increased among cases 8 months before diagnosis, with maximum discrimination between cases and controls 8 months before diagnosis (Fisher exact odds ratio [OR] 2.9, 95% CI 2.1-4.1). An increase in indigestion medication purchases was detected up to 9 months before diagnosis (adjusted conditional logistic regression OR 1.38, 95% CI 1.04-1.83). The ROC analysis for indigestion medication purchases showed a maximum area under the curve (AUC) at 13 months before diagnosis (AUC=0.65, 95% CI 0.57-0.73), which further improved when stratified to late-stage ovarian cancer (AUC=0.68, 95% CI 0.59-0.78).ConclusionsThere is a difference in purchases of pain and indigestion medications among women with and without ovarian cancer up to 8 months before diagnosis. Facilitating earlier presentation among those who self-care for symptoms using this novel data source could improve ovarian cancer patients' options for treatment and improve survival.Trial registrationClinicalTrials.gov NCT03994653; https://clinicaltrials.gov/ct2/show/NCT03994653.

Dataset Information

Using machine learning to investigate self-medication purchasing in England via high street retailer loyalty card data.

Publications

Using machine learning to investigate self-medication purchasing in England via high street retailer loyalty card data.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets