Hypertension identification using inpatient clinical notes from electronic medical records: an explainable, data-driven algorithm study

Prompt engineering for digital mental health: a short review

Prompt engineering, the process of arranging input or prompts given to a large language model to guide it in producing desired outputs, is an emerging field of research that shapes how these models understand tasks, process information, and generate responses in a wide range of natural language processing (NLP) applications. Digital mental health, on the other hand, is becoming increasingly important for several reasons including early detection and intervention, and to mitigate limited availability of highly skilled medical staff for clinical diagnosis. This short review outlines the latest advances in prompt engineering in the field of NLP for digital mental health. To our knowledge, this review is the first attempt to discuss the latest prompt engineering types, methods, and tasks that are used in digital mental health applications. We discuss three types of digital mental health tasks: classification, generation, and question answering. To conclude, we discuss the challenges, limitations, ethical considerations, and future directions in prompt engineering for digital mental health. We believe that this short review contributes a useful point of departure for future research in prompt engineering for digital mental health.

Hypertension in Pregnancy: Current Challenges and Future Opportunities for Surveillance and Research

Hypertension in pregnancy (HP) includes eclampsia/preeclampsia, chronic hypertension, superimposed preeclampsia, and gestational hypertension. In the United States, HP prevalence doubled over the last three decades, based on birth certificate data. In 2019, the estimated percent of births with a history of HP varied from 10.1% to 15.9% for birth certificate data and hospital discharge records, respectively. The use of electronic medical records may result in identifying an additional third to half of undiagnosed cases of HP. Individuals with gestational hypertension or preeclampsia are at 3.5 times higher risk of progressing to chronic hypertension and from 1.7 to 2.8 times higher risk of developing cardiovascular disease (CVD) after childbirth compared with individuals without these conditions. Interventions to identify and address CVD risk factors among individuals with HP are most effective if started during the first 6 weeks postpartum and implemented during the first year after childbirth. Providing access to affordable health care during the first 12 months after delivery may ensure healthy longevity for individuals with HP. Average attendance rates for postpartum visits in the United States are 72.1%, but the rates vary significantly (from 24.9% to 96.5%). Moreover, even among individuals with CVD risk factors who attend postpartum visits, approximately 40% do not receive counseling on a healthy lifestyle. In the United States, as of the end of September 2023, 38 states and the District of Columbia have extended Medicaid coverage eligibility, eight states plan to implement it, and two states proposed a limited coverage extension from 2 to 12 months after childbirth. Currently, data gaps exist in national health surveillance and health systems to identify and monitor HP. Using multiple data sources, incorporating electronic medical record data algorithms, and standardizing data definitions can improve surveillance, provide opportunities to better track progress, and may help in developing targeted policy recommendations.

Exploring the reliability of inpatient EMR algorithms for diabetes identification

INTRODUCTION

Accurate identification of medical conditions within a real-time inpatient setting is crucial for health systems. Current inpatient comorbidity algorithms rely on integrating various sources of administrative data, but at times, there is a considerable lag in obtaining and linking these data. Our study objective was to develop electronic medical records (EMR) data-based inpatient diabetes phenotyping algorithms.

MATERIALS AND METHODS

A chart review on 3040 individuals was completed, and 583 had diabetes. We linked EMR data on these individuals to the International Classification of Disease (ICD) administrative databases. The following EMR-data-based diabetes algorithms were developed: (1) laboratory data, (2) medication data, (3) laboratory and medications data, (4) diabetes concept keywords and (5) diabetes free-text algorithm. Combined algorithms used or statements between the above algorithms. Algorithm performances were measured using chart review as a gold standard. We determined the best-performing algorithm as the one that showed the high performance of sensitivity (SN), and positive predictive value (PPV).

RESULTS

The algorithms tested generally performed well: ICD-coded data, SN 0.84, specificity (SP) 0.98, PPV 0.93 and negative predictive value (NPV) 0.96; medication and laboratory algorithm, SN 0.90, SP 0.95, PPV 0.80 and NPV 0.97; all document types algorithm, SN 0.95, SP 0.98, PPV 0.94 and NPV 0.99.

DISCUSSION

Free-text data-based diabetes algorithm can yield comparable or superior performance to a commonly used ICD-coded algorithm and could supplement existing methods. These types of inpatient EMR-based algorithms for case identification may become a key method for timely resource planning and care delivery.

Filling gaps in female gout: a cross-sectional study of comorbidities in 192 037 hospitalised patients

OBJECTIVE

There is room for improvement in the knowledge of female gout, often noted at risk of gender blindness. This study aims to compare the prevalence of comorbidities in women versus men hospitalised with gout in Spain.

METHODS

This is an observational, multicentre, cross-sectional study in public and private Spanish hospitals analysing the minimum basic data set from 192 037 hospitalisations in people with gout (International Classification of Diseases, Ninth Revision (ICD-9) coding) from 2005 to 2015. Age and several comorbidities (ICD-9) were compared by sex, with a subsequent stratification of comorbidities by age group. The association between each comorbidity and sex was assessed using multivariable logistic regression. A clinical decision tree algorithm was constructed to predict the sex of patients with gout based on age and comorbidities alone.

RESULTS

Women with gout (17.4% of the sample) were significantly older than men (73.9±13.7 years vs 64.0±14.4 years, p<0.001). Obesity, dyslipidaemia, chronic kidney disease, diabetes mellitus, heart failure, dementia, urinary tract infection and concurrent rheumatic disease were more common in women. Female sex was strongly associated with increasing age, heart failure, obesity, urinary tract infection and diabetes mellitus, while male sex was associated with obstructive respiratory diseases, coronary disease and peripheral vascular disease. The decision tree algorithm built showed an accuracy of 74.4%.

CONCLUSIONS

A nationwide analysis of inpatients with gout in 2005-2015 confirms a different comorbidity profile between men and women. A different approach to female gout is needed to reduce gender blindness.