Computerized Physician Order Entry Research Articles

A multifaceted quality improvement intervention on venous thromboembolism prophylaxis compliance in hospitalized medical patients at a comprehensive cancer center.

Previous studies suggest that quality improvement initiatives focused on hospital-acquired venous thromboembolism have a positive impact on prescribing rates of venous thromboembolism prophylaxis, especially those that incorporate computerized changes. We conducted a quality improvement project to determine whether education and computerized prescriber order entry system changes affect venous thromboembolism prophylaxis compliance rates in hospitalized medical patients at a Comprehensive Cancer Center. Between 1 January 2021 and 31 January 2023, 37,739 non-surgical, adult patient encounters with a length of stay > 48 h were analyzed in our study. From 18 December 2021 to 8 March 2022, provider education was delivered to the three largest admitting services, and computerized prescriber order entry changes were implemented incorporating a mandatory requirement to either order venous thromboembolism prophylaxis or document a contraindication for all patients at moderate venous thromboembolism risk. Monthly venous thromboembolism prophylaxis compliance rates, as defined by the Centers for Medicare and Medicaid Services VTE-1 metric, increased from a mean of 74% to 93% after the interventions. This change was driven primarily by an increased utilization of mechanical venous thromboembolism prophylaxis from 37% to 53%. Our study demonstrated that a multi-faceted intervention incorporating provider education and computerized prescriber order entry system changes can significantly increase venous thromboembolism prophylaxis compliance rates in cancer patients.

P‐PB‐11 | Implementation of a Computerized Physician Order Entry Alert for Double Dose Red Cell Orders

P‐PB‐11 | Implementation of a Computerized Physician Order Entry Alert for Double Dose Red Cell Orders

Unveiling medication errors in liver transplant patients towards enhancing the imperative patient safety

BackgroundMedication errors (MEs) are a significant healthcare problem that can harm patients and increase healthcare expenses. Being immunocompromised, liver-transplant patients are at high risk for complications if MEs inflict harmful or damaging effects. The present study reviewed and analyzed all MEs reported in Liver Transplant Patients. MethodsAll MEs in the Liver Transplant Patients admitted between January 2016 to August 2022 were retrieved through the computerized physician order entry system, which two expert pharmacists classified according to the type and severity risk index. ResultsA total of 314 records containing 407 MEs were committed by at least 71 physicians. Most of these errors involved drugs unrelated to managing liver-transplant-related issues. Antibiotic prescriptions had the highest mistake rate (17.0%), whereas immunosuppressants, routinely used in liver transplant patients, rank second with fewer than 14% of the identified MEs. The most often reported MEs (43.2%) are type-C errors, which, despite reaching patients, did not cause patient harm. Subgroup analysis revealed several factors associated with a statistically significant great incidence of MEs among physicians treating liver transplant patients. ConclusionAlthough a substantial number of MEs occurred with liver transplant patients, the majority are not related to liver-transplant medications, which mainly belonged to type-C errors. This could be attributed to polypharmacy of transplant patients or the heavy workload on health care practitioners. Improving patient safety requires adopting regulations and strategies to promptly identify MEs and address potential errors.

Implementation of a clinical decision support system for the optimization of antidiabetic drug orders by pharmacists.

The objective of the study was to describe the impact of a clinical decision support system (CDSS) on antidiabetic drug management by clinical pharmacists for hospitalized patients with T2DM. We performed a retrospective, single-centre study in a teaching hospital, where clinical pharmacists analysed prescriptions and issued pharmacist interventions (PIs) through a computerized physician order entry (CPOE) system. A CDSS was integrated into the pharmacists' workflow in July 2019. We analysed PIs during 2 periods of interest: one before the introduction of the CDSS (from November 2018 to April 2019, PIs issued through the CPOE alone) and one afterwards (from November 2020 to April 2021, PIs issued through the CPOE and/or the CDSS). The study covered nondiabetology wards as endocrinology, diabetes and metabolism departments were not computerized at the time of the study. There were 203 PIs related to antidiabetic drugs in period 1 and 319 in period 2 (a 57.5% increase). Sixty-four of the 319 PIs were generated by the CDSS. Noncompliance/contraindication was the main problem identified by the CDSS (41 PIs, 68.4%), and 57.8% led to discontinuation of the drug. Most of the PIs issued through the CDSS corresponded to orders that had not been flagged up by clinical pharmacists using the CPOE. Conversely, most alerts about indications that were not being treated were detected by the clinical pharmacists using the CPOE and not by the CDSS. Use of CDSS by clinical pharmacists improved antidiabetic drug management for hospitalized patients with T2DM. The CDSS might add value to diabetes care in nondiabetology wards by decreasing the frequency of potentially inappropriate prescriptions and adverse drug reactions.

Do junior doctors make more prescribing errors than experienced doctors when prescribing electronically using a computerised physician order entry system combined with a clinical decision support system? A cross-sectional study

ObjectivesPrescribing errors can lead to inconvenience, morbidity and mortality. It is therefore crucial to educate doctors to prescribe safely, efficiently and effectively. To create an effective educational programme, it is...

Artificial intelligence in the pre-analytical phase: State-of-the art and future perspectives.

The use of artificial intelligence (AI) has become widespread in many areas of science and medicine, including laboratory medicine. Although it seems obvious that the analytical and post-analytical phases could be the most important fields of application in laboratory medicine, a kaleidoscope of new opportunities has emerged to extend the benefits of AI to many manual labor-intensive activities belonging to the pre-analytical phase, which are inherently characterized by enhanced vulnerability and higher risk of errors. These potential applications involve increasing the appropriateness of test prescription (with computerized physician order entry or demand management tools), improved specimen collection (using active patient recognition, automated specimen labeling, vein recognition and blood collection assistance, along with automated blood drawing), more efficient sample transportation (facilitated by the use of pneumatic transport systems or drones, and monitored with smart blood tubes or data loggers), systematic evaluation of sample quality (by measuring serum indices, fill volume or for detecting sample clotting), as well as error detection and analysis. Therefore, this opinion paper aims to discuss the state-of-the-art and some future possibilities of AI in the preanalytical phase.

Trends in computerized provider order entry: 20-year bibliometric overview.

Drug-related problems (DRPs) can lead to serious health issues and have significant economic impacts on healthcare systems. One solution to address this issue is the use of computerized physician order entry systems (CPOE), which can help prevent DRPs by reducing the risk of medication errors. The purpose of this study is to provide an analysis on scientific production of the past 20 years in order to describe trends in academic publishing on CPOE and to identify the major topics as well as the predominant actors (journals, countries) involved in this field. A PubMed search was carried out to extract articles related to computerized provider order entry during the period January 1st 2003- December 31st 2022 using a specific query. Data were downloaded from PubMed in Extensible Markup Language (XML) and were processed through a dedicated parser. A total of 2,946 articles were retrieved among 623 journals. One third of these articles were published in eight journals. Publications grew strongly from 2002 to 2006, with a dip in 2008 followed by an increase again in 2009. After 2009, there follows a decreasing until 2022.The most producing countries are the USA with 51.39% of the publication over the period by France (3.80%), and Canada (3.77%). About disciplines, the top 3 is: "medical informatics" (21.62% of articles), "pharmacy" (19.04%), and "pediatrics" (6.56%). This study provides an overview of publication trends related to CPOE, which exhibited a significant increase in the first decade of the 21st century followed by a decline after 2009. Possible reasons for this decline include the emergence of digital health tools beyond CPOE, as well as healthcare professionals experiencing alert fatigue of the current system. Future research should focus on analyzing publication trends in the field of medical informatics and decision-making tools to identify other areas of interest that may have surpassed the development of CPOE.

Design, development, implementation, and evaluation of a severe drug–drug interaction alert system in the ICU: An analysis of acceptance and override rates

Background and objectiveThe override rate of Drug-Drug Interaction Alerts (DDIA) in Intensive Care Units (ICUs) is very high. Therefore, this study aimed to design, develop, implement, and evaluate a severe Drug-Drug Alert System (DDIAS) in a system of ICUs and measure the override rate of this system. MethodsThis is a cross-sectional study that details the design, development, implementation, and evaluation of a DDIAS for severe interactions into a Computerized Provider Order Entry (CPOE) system in the ICUs of Nemazee general teaching hospitals in 2021. The patients exposed to the volume of DDIAS, acceptance and overridden of DDIAS, and usability of DDIAS have been collected. The study was approved by the local Institutional Review Board (IRB) and; the ethics committee of Shiraz University of Medical Science on date: 2019-11-23 (Approval ID: IR.SUMS.REC.1398.1046). ResultsThe knowledge base of the DDIAS contains 9,809 severe potential drug-drug interactions (pDDIs). A total of 2672 medications were prescribed in the population study. The volume and acceptance rate for the DDIAS were 81 % and 97.5 %, respectively. The override rate was 2.5 %. The mean System Usability Scale (SUS) score of the DDIAS was 75. ConclusionThis study demonstrates that implementing high-risk DDIAS at the point of prescribing in ICUs improves adherence to alerts. In addition, the usability of the DDIAS was reasonable. Further studies are needed to investigate the establishment of severe DDIAS and measure the prescribers’ response to DDIAS on a larger scale.

Using alert dwell time to filter universal clinical alerts: A machine learning approach

BackgroundAlerts in computerized physician order entry (CPOE) systems can improve patient safety. However, alerts in rule-based systems cannot be customized based on individual patient or user characteristics. This limitation can lead to the presentation of irrelevant alerts and subsequent alert fatigue. ObjectiveWe used machine learning approaches with alert dwell time to filter out irrelevant alerts for physicians based on contextual factors. MethodsWe utilized five machine learning algorithms and a total of 1,120 features grouped into six categories: alert, demographic, environment, diagnosis, prescription, and laboratory results. The output of the models was the alert dwell time within a specified time window to determine the optimal range by the sensitivity analysis. ResultsWe used 813,026 records (19 categories) from the hospital's outpatient clinic data from 2020 to 2021. The sensitivity analysis showed that a time window with a range of 0.3–4.0 s had the best performance, with an area under the receiver operating characteristic (AUROC) curve of 0.73 and an area under the precision–recall curve (AUPRC) of 0.97. The model built with alert and demographic feature groups showed the best performance, with an AUROC of 0.73. The most significant individual feature groups were alert and demographic, with AUROCs of 0.66 and 0.62, respectively. ConclusionOur study found that alerts and user and patient demographic features are more crucial than clinical features when constructing universal context-aware alerts. Using alert dwell time in combination with a time window is an effective way to determine the trigger status of an alert. The findings of this study can provide useful insights for researchers working on specific and universal context-aware alerts.

Traduction et adaptation au contexte hospitalier français d’un outil de codification des erreurs de prescription liées aux logiciels

Prescribing errors related to computerized physician order entry are current and may have serious consequences for patients. They can be detected by pharmacists during prescriptions analysis and lead to pharmacist's interventions. In France, few monocentric studies have studied Pharmacist Interventions triggered by prescribing errors identified as System-Related Errors (PISREs) in French hospitals. However, their respective analysis method prevent any comparison between computerized physician order entry systems in order to identify the safest and rule out the most dangerous. A computerized physician prescribing error related to the software is characterized by its causes, consequences and mechanism of occurrence. US researchers have developed and validated a tool to classify and illustrate these three characteristics. The objectives of this article are to present this tool, to propose a French adaptation and to describe the perspectives analyze and understand prescription errors related to computerized physician order entry based on data of Act-IP©. The adaptation was performed using PISREs extracted from the Act-IP© observatory of the French Society of Clinical Pharmacy. Each item of the codification is illustrated with an example of PI. We are considering a training plan in order to allow wide use of this tool. Once adopted this tool, the next step will be to organize a prospective multicenter study including as many computerized prescription order entry systems as possible. The aim of this study will be identifying the safest systems. Consequently, it will then be possible to have arguments to qualify the most dangerous and thus propose their withdrawal from the market.

Prescribing Practices at a Residential Treatment Center – from Paper to Plastic (Computers)

ABSTRACT Medication prescription procedures were reviewed at psychiatric residential treatment centers (RTC), and a quality improvement project was conducted in order to implement integrated electronic medical records (EMR) and computerized provider order entry (CPOE). To assess the impact of preexisting prescribing practices on RTCs, surveys were sent anonymously to current and past providers of regional and local RTCs. The surveys assessed the impact of the prescribing protocol on workflow, the frequency of medication errors, the impact such errors had on patient care, the overall provider satisfaction, and the discrepancies between electronic and paper charts. Current providers were then re-surveyed after implementing the integrated CPOE system in May 2020, during the COVID-19 pandemic. The first round of provider survey results revealed that the majority of staff (n = 10, 70%) observed frequent medication errors and were largely dissatisfied with the current medication ordering and administration process (n = 10; 70%). Incident Response Improvement System (IRIS) reports revealed the most common medication errors were omissions, incorrect dosing, and incorrect medications administered. After the initiation and implementation of the CPOE, staff satisfaction increased with the process (n = 14; 50%) noting that medication errors were less frequent than before (n = 14; 50%).

OVERVIEW OF MEDICATION ERROR INCIDENCE IN HOSPITALS IN VARIOUS COUNTRIES: LITERATURE REVIEW

Background: Medication error is one of the most common types of errors contributing to patient safety incidents in hospitals. In addition to their numerous cases and high costs, medication errors also contribute to deaths in various countries. Aims: This study describes the incidence of medication errors in hospitals in various countries, determines the phase of errors that occurred the most, and describes preventive strategies for medication errors in hospitals. Methods: This study was conducted between April 2021-July 2021 using the literature review method. Data were retrieved from ProQuest, ScienceDirect, PubMed, Google Scholar, and Garuda RistekBRIN. The data are generalized and extracted in a table based on the incidence of medication errors and preventive strategies. Results: Of the eleven included studies, the incidence of medication errors in one study conducted in Nigeria was the highest (80%). Four of the eleven studies were conducted in India with varying incidence rates. Most errors occurred during the prescribing stage. The number of reports and the number of events actually have no relevance. It can be said that countries with a high number of reporting have good reporting indicators. Strategies for preventing medication errors include the implementation of an information system in the CPOE (Computerized Physician Order Entry) form and providing training for staff. Conclusion: The difference in the incidence of medication errors in developing and developed economies can be attributed to factors in the healthcare system and the lower prescribing ratio and nurse ratio in developing countries. Keywords: literature review, medication errors, patient safety

OVERVIEW OF MEDICATION ERROR INCIDENCE IN HOSPITALS IN VARIOUS COUNTRIES: LITERATURE REVIEW

Background: Medication error is one of the most common types of errors contributing to patient safety incidents in hospitals. In addition to their numerous cases and high costs, medication errors also contribute to deaths in various countries. Aims: This study describes the incidence of medication errors in hospitals in various countries, determines the phase of errors that occurred the most, and describes preventive strategies for medication errors in hospitals. Methods: This study was conducted between April 2021-July 2021 using the literature review method. Data were retrieved from ProQuest, ScienceDirect, PubMed, Google Scholar, and Garuda RistekBRIN. The data are generalized and extracted in a table based on the incidence of medication errors and preventive strategies. Results: Of the eleven included studies, the incidence of medication errors in one study conducted in Nigeria was the highest (80%). Four of the eleven studies were conducted in India with varying incidence rates. Most errors occurred during the prescribing stage. The number of reports and the number of events actually have no relevance. It can be said that countries with a high number of reporting have good reporting indicators. Strategies for preventing medication errors include the implementation of an information system in the CPOE (Computerized Physician Order Entry) form and providing training for staff. Conclusion: The difference in the incidence of medication errors in developing and developed economies can be attributed to factors in the healthcare system and the lower prescribing ratio and nurse ratio in developing countries. Keywords: literature review, medication errors, patient safety

Patient Safety of Perioperative Medication Through the Lens of Digital Health and Artificial Intelligence.

Perioperative medication has made significant contributions to enhancing patient safety. Nevertheless, administering medication during this period still poses considerable safety concerns, with many errors being detected only after causing significant physiological disturbances. The intricacy of medication administration in the perioperative setting poses specific challenges to patient safety. To address these challenges, implementing potential strategies and interventions is critical. One such strategy is raising awareness and revising educational curricula regarding drug safety in the operating room. Another crucial strategy is recognizing the importance of redundancy and multiple checks in the operating room as a hallmark of medication safety, which is not a common practice. Digital health technologies and artificial intelligence (AI) also offer the potential to improve perioperative medication safety. Computerized physician order entry systems, electronic medication administration records, and barcode medication administration systems have been proven to reduce medication errors and improve patient safety. By implementing these strategies and interventions, health care professionals can enhance the safety of perioperative medication administration and improve patient outcomes.

Impact of a computerized physician order entry system on medication safety in pediatrics-The AVOID study.

One of the most critical steps in the medication process on pediatric wards is the medical prescription. This study aims to investigate the impact of a computerized physician order entry (CPOE) system on Adverse Drug Events (ADEs) and potentially harmful ADEs (pot ADEs) in comparison with paper-based documentation in a general pediatric ward at a German University hospital. A prospective pre-post study was conducted. All patients aged 17 years or younger were observed during the study periods (5 months pre- and postimplementation). Issues Regarding Medication (IRM) were identified by intensive chart review. Events were assessed regarding causality (WHO), severity (WHO; Dean & Barber for MEs), and preventability (Shumock) and classified into (pot) ADEs, (pot) Medication errors (ME), Adverse drug Reactions (ADR), and Other incidents (OI) accordingly. Total of 333 patients with medication were included in the paper-based prescribing cohort (phase I) and 320 patients with medication in the electronic prescribing cohort (phase II). In each cohort, patients received a median number of four different drugs (IQR 5 and IQR 4). A total of 3966 IRM was observed. During the hospitalization, 2.7% (n = 9) patients in phase I and 2.8% (n = 9) in phase II experienced an ADE. Potentially harmful MEs were less often observed in the cohort with electronic prescribing (n = 228 vs. n = 562). The mean number per patient significantly decreased from 1.69 to 0.71 (p < .01). The implementation of a CPOE system resulted in a reduction of issues regarding medication, particularly MEs with the potential to harm patients decreased significantly.

Computerized physician order entry (CPOE) evaluation – what can we measure and how can we do it?

Computerized physician order entry (CPOE) evaluation – what can we measure and how can we do it?

Pseudorandomized Testing of a Discharge Medication Alert to Reduce Free-Text Prescribing.

Pseudorandomized testing can be applied to perform rigorous yet practical evaluations of clinical decision support tools. We apply this methodology to an interruptive alert aimed at reducing free-text prescriptions. Using free-text instead of structured computerized provider order entry elements can cause medication errors and inequity in care by bypassing medication-based clinical decision support tools and hindering automated translation of prescription instructions. The objective of this study is to evaluate the effectiveness of an interruptive alert at reducing free-text prescriptions via pseudorandomized testing using native electronic health records (EHR) functionality. Two versions of an EHR alert triggered when a provider attempted to sign a discharge free-text prescription. The visible version displayed an interruptive alert to the user, and a silent version triggered in the background, serving as a control. Providers were assigned to the visible and silent arms based on even/odd EHR provider IDs. The proportion of encounters with a free-text prescription was calculated across the groups. Alert trigger rates were compared in process control charts. Free-text prescriptions were analyzed to identify prescribing patterns. Over the 28-week study period, 143 providers triggered 695 alerts (345 visible and 350 silent). The proportions of encounters with free-text prescriptions were 83% (266/320) and 90% (273/303) in the intervention and control groups, respectively (p = 0.01). For the active alert, median time to action was 31 seconds. Alert trigger rates between groups were similar over time. Ibuprofen, oxycodone, steroid tapers, and oncology-related prescriptions accounted for most free-text prescriptions. A majority of these prescriptions originated from user preference lists. An interruptive alert was associated with a modest reduction in free-text prescriptions. Furthermore, the majority of these prescriptions could have been reproduced using structured order entry fields. Targeting user preference lists shows promise for future intervention.

Machine Learning Approach for Prediction of the Test Results of Gonadotropin-Releasing Hormone Stimulation: Model Building and Implementation.

Precocious puberty in girls is defined as the onset of pubertal changes before 8 years of age, and gonadotropin-releasing hormone (GnRH) agonist treatment is available for central precocious puberty (CPP). The gold standard for diagnosing CPP is the GnRH stimulation test. However, the GnRH stimulation test is time-consuming, costly, and requires repeated blood sampling. We aimed to develop an artificial intelligence (AI) prediction model to assist pediatric endocrinologists in decision making regarding the optimal timing to perform the GnRH stimulation test. We reviewed the medical charts of 161 girls who received the GnRH stimulation test from 1 August 2010 to 31 August 2021, and we selected 15 clinically relevant features for machine learning modeling. We chose the models with the highest area under the receiver operating characteristic curve (AUC) to integrate into our computerized physician order entry (CPOE) system. The AUC values for the CPP diagnosis prediction model (LH ≥ 5 IU/L) were 0.884 with logistic regression, 0.912 with random forest, 0.942 with LightGBM, and 0.942 with XGBoost. For the Taiwan National Health Insurance treatment coverage prediction model (LH ≥ 10 IU/L), the AUC values were 0.909, 0.941, 0.934, and 0.881, respectively. In conclusion, our AI predictive system can assist pediatric endocrinologists when they are deciding whether a girl with suspected CPP should receive a GnRH stimulation test. With proper use, this prediction model may possibly avoid unnecessary invasive blood sampling for GnRH stimulation tests.

Impact of clinical pharmacist interventions in a bone and joint infection orthoseptic surgery unit

To assess the impact of interventions of a clinical pharmacist in a unit of orthopedic surgery specialized in bone and joint infections. Daily, in routine, a clinical pharmacist analyzed medication prescribed to inpatients via a computerized physician order entry (CPOE) (Phedra software). His attention was particularly focused on the impact of antibiotics on other medications. For this study, all of the pharmacist interventions (PI) have been retrospectively collected, then anonymized, and assessed over a two-month period. Thirty-eight patients were hospitalized during the study period, with a mean age of 63 years old. Forty-five interventions were identified which represents a mean of 1.18 pharmaceutical interventions per patient. Most of them concerned lack of follow-up (24%) and drug-drug interactions (22%) and widely non-anti-infectious medication (35 interventions) with levothyroxine (10 interventions) as the most involved non-anti-infectious molecule. Among antibiotics, with respectively 9 and 8 interventions, rifampicin and fluoroquinolones (6 interventions for moxifloxacin) were the most concerned notably for drug-drug interactions with usual treatment. In this observational retrospective study, 1.18 pharmacist interventions (PI) per patient were observed. Most of them are lack of follow-up and drug-drug interactions especially with usual treatment of patients. Moxifloxacin and rifampicin were the most antibiotics involved. Patients' characteristics (older, polypharmacy), long-term hospitalization and surgery are known to be predictive factors of medication errors and this study highlights the importance of the presence of clinical pharmacist in orthopedic surgery wards.

Developing a minimum data set for cardiovascular Computerized Physician Order Entry (CPOE) and mapping the data set to FHIR: A multi-method approach

Many medical errors occur in the process of treating cardiovascular patients, and most of these errors are related to prescription errors. There are several, one of the methods to prevent prescription errors is the use of a computerized physician order entry (CPOE) system. One of the obstacles of implementing this system is improper design and non-compliance with user needs. one of the issues that should be considered in designing information systems is having a standard minimum data set (MDS). Although many computerized physicians order entry (CPOE) systems have been developed in the world, no study has identified the necessary data and minimum data set (MDS) of CPOE system, and published the process of creating this MDS. This study aimed to develop an MDS for cardiovascular CPOE and standardize it with Fast Healthcare Interoperability Resources (FHIR). A multi-method approach including systematic review for identifying data elements of CPOE, reviewing the content of medical records, validation of the data elements using the expert panel and, determination of the necessary data elements using a survey was conducted. Classification of the data elements and mapping them to FHIR were done to facilitate data sharing and integration with the electronic health record (EHR) system as well as to reduce data diversity. The final data elements of MDS were categorized into 5 main categories of FHIR (foundation, base, clinical, financial, and specialized) and 146 resources, where possible. Mapping was done by one of the researchers and checked and verified by the second researcher. Non-mapped data elements were added to relevant resources as extensions of existing FHIR resources. In total, 270 data elements were identified from the systematic review. After reviewing the content of 20 patients' medical records, 28 data elements were identified. After combination of data elements of two previous phases and removing duplication, 282 data elements remained. Data elements that were considered necessary to be included in CPOE by conducting a survey among cardiovascular physicians were 109 elements. From 146 resources of FHIR, the data elements of this MDS are covered by 5 resources. This study introduced an MDS for cardiovascular CPOE by combining suggested data elements of previous research, and the practical and local requirements identified in patients' medical records. To facilitate data sharing and integration with EHR, reduce data diversity, and also to categorize data, this MDS was standardized with FHIR. The steps we used to develop this MDS could be a model for creating MDS in other CPOEs and health information systems. This is the first time that the process of developing an MDS for cardiovascular CPOE has been presented in the literature.