EHR data and AI to predict response to antidepressant treatment

Antidepressants are frequently prescribed for adults with depression, a common and often disabling psychiatric condition. However, identifying the most effective treatment for a particular patient is often a trial-and-error process that can result in prolonged morbidity, disability, and exposure to adverse effects, as well as substantial healthcare costs. Precision psychiatry aims to optimise treatment matching using patient-specific profiles, but there are few evidence-based predictors available to clinicians initiating antidepressant treatment.

Although average response rates are similar across different antidepressant classes, individual responses can vary widely in clinical practice. Therefore, accurately and scalably guiding antidepressant selection presents specific challenges. The gold standard for characterising antidepressant response from electronic health records (EHRs) remains expert chart review, which is labor- and time-intensive.

However, advances in machine learning (ML) and the growing availability of large-scale health data, such as EHRs, offer new opportunities for developing clinical decision-support tools that may address this challenge. In a recent study published in the peer-reviewed open-access medical journal Nature Partner Journals (npc) Digital Medicine, researchers used machine learning models to accurately predict differential treatment response probabilities for patients and between antidepressant classes based on real-world EHR data. The pipeline incorporated AI and non-AI features, as well as unstructured data (i.e. clinical notes) to maximize the use of information contained in EHRs.

The study included 17,556 patients who received a new antidepressant prescription from non-psychiatrists, and data were obtained from 20 years of EHRs spanning from January 1990 to August 2018. The patients had at least one International Classification of Diseases (ICD) code for depression and at least one ICD code for non-recurrent depression during their history.

ICD codes from EHR data were obtained for adult patients (age ≥ 18 years) with at least one visit (the first visit with an antidepressant prescription is defined as the “index visit” for each patient) with a diagnostic ICD code for a depressive disorder (defined as ICD-9-CM: 296.20–6, 296.30–6, and 311; ICD-10-CM: F32.0–9, F33.0–9) co-occurring with an antidepressant prescription, and at least one ICD code for non-recurrent depression (ICD-9-CM: 296.20–6 and 311; ICD-10-CM: F32.0–9) any time during their history.

The resulting models achieved good accuracy, discrimination, and positive predictive value, which could be valuable for further efforts aiming to provide clinical decision support for prescribers. However, the researchers noted several limitations, including missing data in EHRs(e.g. patients who may receive some of their care outside of the healthcare system), and secular trends in clinician prescribing or documentation practices that may have affected model performance.

In summary, the study presents a novel computational pipeline based on real-world EHR data for predicting differential responses to commonly used classes of antidepressants. The approach demonstrated here could be adapted to a wide variety of other clinical applications for optimising and individualising treatment selection.

REFERENCES:

  1. Sheu, Yh., Magdamo, C., Miller, M. et al. AI-assisted prediction of differential response to antidepressant classes using electronic health records. npj Digit. Med. 6, 73 (2023). https://doi.org/10.1038/s41746-023-00817-8


International Pi Day 2023

As it is Pi Day 2023, I like to dedicate this post to relate to the use of pi in Health Information Management(HIM).

We all first met the pi in elementary geometry at school and it occurs in one of the first equations that you can solve; that is to find the circumference or areas of circles.

We remember that Pi (π) is a mathematical constant that represents the ratio of the circumference of a circle to its diameter, approximately equal to 3.14159—when for pi, you substitute the fraction 22/7. Although it is a simple ratio of the circumference of a circle to its diameter, the decimal places appear to go on forever.

Pi is used extensively in mathematics and science, and its calculation has applications in various fields in healthcare too, including HIM.

For this post, I can think of three applications of pi, when pi is integral to HIM.

While the electronic medical record (EMR) is the core informational system for patient management across the healthcare system, the radiology information system (RIS) is considered the core system for the electronic management of imaging departments.

The RIS is the first example I can think of from my past inter-professional collaboration in a hospital setting, that pi is related in one way to HIM, i.e. through its use in medical imaging.

Medical imaging technologies, such as computed tomography (CT) and magnetic resonance imaging (MRI), use pi in their calculations to generate accurate images of the human body. The measurements of these images are based on the circumference and diameter of the scanned area, and pi is used to calculate these measurements accurately.

In addition to the RIS, pi is used in the calculation of various health-related metrics, such as body mass index (BMI) and blood pressure.

Obesity is a common problem worldwide that independently confers risk for chronic disease and early mortality.

HIM professionals use the International Classification of Diseases, Tenth Revision (ICD-10) codes as a tool for medical diagnoses, to medically code obesity, and also to identify obesity documentation in the electronic medical records(EMR) problem list.

Other available sources in the EMR use an individual’s body mass index (BMI) to identify and categorise obesity into the Z codes of ICD-10.

The ability to identify and manage the care of patients who meet the criteria for obesity in ambulatory settings has significantly improved with the increased use of health information technology, especially with the use of EMRs.

It is here in the EMR, that It is worthy to take note of the humble pie in the calculation of BMI involves dividing a person’s weight in kilograms by the square of their height in meters, which involves the use of pi in the calculation of the area of a circle (since the formula for the area of a circle is A = πr^2).

As HIM professionals involved in research, we surely used pi in statistical analysis and modeling in HIM.

For example, pi is used in the calculation of confidence intervals, which are used to estimate the range of values within which a population parameter (such as a mean or proportion) is likely to fall. This is important in HIM, as it allows researchers and analysts to make inferences about the health of a population based on a sample of data.

In summary, pi is used in various ways in HIM, including medical imaging, the calculation of health-related metrics, and statistical analysis and modeling.

Its precise calculation and properties make it a valuable tool in the field of HIM.

Moving to ICD-11

The International Classification of Diseases(ICD) is the international standard for the systematic recording, reporting, analysis, interpretation, and comparison of mortality and morbidity data.

The World Health Organisation (WHO) presented and released the 11th edition of ICD(ICD-11) at the World Health Assembly on May 25, 2019, for adoption by member states,

This release has since come into effect on January 1, 2022, to replace the 10th revision(ICD-10), currently in use.

While ICD-10 is still widely used, unfortunately, despite the updating process, ICD-10 is known to be clinically outdated, and structural changes are needed in some chapters.

There is also an increasing need to operate in an electronic environment, as well as the need to capture more information for morbidity-use cases.

According to the WHO, the 11th revision is a scientifically rigorous product that accurately reflects contemporary health and medical practice and represents a significant upgrade from earlier revisions.

The WHO ICD-11 revision goals include to;
1:
Ensure that ICD-11 will function in an electronic environment by:
a. presenting a digital product
b. providing linkage with terminologies (e.g., SNOMED)
c. defining ICD Categories by “logical operational rules” on their associations and details
d. supporting electronic health records & information systems;

2:
Provide a multi-purpose and coherent classification for mortality, morbidity, primary care, clinical care, research, and public health;

3:
Consistency & interoperability across different uses; and

4:
Deliver an international, multilingual reference standard for scientific comparability, i.e. in English, French, Spanish, Russian, Chinese, and Arabic.

Analysis of mortality and cause of death data using ANACoD3

During a part of the continuing WHO ICD-11 webinar series,  in collaboration with the Surveys, CRVS, & Health Service Data Unit, the Classifications and Terminologies Unit of the World Health Organisation (WHO), launched the Analysing Mortality and of Cause of Death 3 (ANACoD3) on September 29, 2021.

ANACoD3 is a new electronic online tool that helps to perform a comprehensive and systematic analysis of mortality and cause of death data.

References: WHO ICD-11 Webinar series – ANACoD3 tool launch, available online{link opens in a new tab of the same window): https://www.who.int/news-room/events/detail/2021/09/29/default-calendar/who-icd-11-webinar-series—anacod3-tool-launch