Improving Mental Health Care With Data
May is Mental Health Awareness Month. It is the time of year when we are (extra) encouraged to donate to the nonprofits who serve one of the most vulnerable of populations. It is time to highlight the topics surrounding mental health, dispelling stigma, and celebrating innovations that are improving lives. In keeping with the spirit of this month, this blog post will give an overview of the mental health care system challenges in America, summarized from publicly available reports and from my own professional experience. I will also explore examples of how technologists and data scientists are seeking to create practical data-driven solutions to some of these complicated problems.
The United States is considered to be a leader in the movement towards “deinstitutionalization”. Beginning in the 1960s, this movement was inspired by the idea that every patient with mental illness would be better cared for in small community settings rather than in large hospitals. U.S. policymakers, propelled by financial incentives and new drug protocols quickly agreed and closed the majority of state-operated psychiatric hospitals.
The U.S. reduced hospitals more drastically than any other Western country. Between 1955 and 1994, roughly 487,000 mentally ill patients were discharged from state hospitals. Currently, around 2.2 million people who are severely mentally ill do not receive any psychiatric treatment at all. About 200,000 of those who suffer from schizophrenia or bipolar disorder are homeless.
The idea of community-based care was truly beautiful in theory because the goal was to give patients more autonomy and to integrate them into society. Unfortunately, it was never fully implemented, especially in small towns or rural communities where it was logistically not feasible. The hospitals closed, but community-based clinics did not replace them. This has led to the current crisis in mental health care in America.
The treatment, stability, and shelter that are so desperately needed, were lost, and the people who needed that the most found their way into other public institutions, mainly jails and prisons. People who can’t get treatment for their illness rotate through jails, shelters or the streets. They crowd emergency rooms, sometimes purposely making themselves physically sick, waiting for hospital beds to open and hoping to get a clean shower and a hot meal. Some people become violent, but more often they are victims of violence.
The remaining state psychiatric hospitals are now underfunded and understaffed, their employees are paid low wages and there is high turnover. The environments are stressful, overcrowded and at times dangerous. This creates an incredibly low quality of care and a cycle of continual trauma. Hospitals are simply overrun, with about 12 beds available per 100,000 people. According to the Treatment Advocacy’s Center’s report, “Deinstitutionalization: A Failed History,” this is the same ratio as it was in 1850.
Private psychiatric hospitals cannot keep up with the patient demand as well, which allows hospital administrators to pick the patients they want to treat, and to refuse to accept people with severe mental illness who they determine will be “unsuccessful” in their program. Yes, that is right — it is legal for private hospitals to refuse medical treatment to patients that are too sick.
In addition to the absence of hospital beds, the lack of community resources, and the ongoing stigma and discrimination, there is a glaring issue with the coordination of care. Hospitals, psychiatrists, primary care doctors, and other specialists often do not communicate with each other or share critical information. This can be attributed to outdated record management systems, inefficient methods to ensure patient information privacy — HIPAA, and frequent changes in medical staff and case management. Clinicians rely on self-reported information or partial medical records to make critical decisions.
In its current state, the American mental health care system often fails to prevent avoidable tragedy and injustice. These consequences not only impact individuals with a mental illness, but they also reverberate through families, communities, schools, workplaces, and society. In order to make meaningful changes, mental health care must be brought into the 21st century.
There is some good news though- scientific research and innovative companies focused on the intersection of automation, data science, and mental health care are growing.
Optimizing and Personalizing interventions for people with schizophrenia Across the Lifespan.
Approximately 2.8 million people live with schizophrenia in America, 40% of them will go untreated. Research has shown that early intervention, including medication, can change the nature of this disease and vastly improve long term outcomes for patients. Antipsychotic medication is a vital component of the treatment of schizophrenia. Early experiences can have a lasting impact on patient attitudes toward medication, and the first-episodes of psychosis is a critical time to optimize treatment.
OPAL is a platform to develop, adapt, and examine interventions that address important issues in the treatment and delivery of services for schizophrenia. They use expertise from Mental Health Data Science, as well as Biostatistics to develop machine learning precision models and data analysis. One project on the horizon is focused on medication management through smartphone data.
Prescribing the best possible medication and dosage can be challenging. A contributing factor to this issue is the lack of accurate information about how the medication is treating the symptoms, the side effects, as well as the corresponding behavioral, cognitive, and emotional experiences of the patient. Psychiatrists typically rely on a patients’ recollection of how they felt over the past month. This is problematic since people diagnosed with schizophrenia are vulnerable to memory difficulties and cognitive biases.
This project will use a smartphone app to collect real-time symptom and current patient functioning data from patients who are experiencing a first-time episode of psychosis. These inputs will allow clinical teams to make better treatment decisions with relevant and timely data. The intention is to treat the illness at the early stages with the best possible information, in order to make a lasting life-long impact.
Data Kind is a nonprofit that brings together volunteer data scientists and social change organizations to “tackle critical humanitarian issues in the fields of education, poverty, health, human rights, the environment, and cities.”
In a joint effort between Data Kind volunteers and the Chicago nonprofit Thresholds, they worked to aggregate patient data across multiple platforms and use predictive analytics to identify high-risk patients who could use early intervention. Without access to important and up-to-date patient data, it is difficult for health care providers to make sound decisions about the care and assistance that patients need.
The team began the project by creating a data warehouse encompassing raw data pulled from Thresholds’ internal databases, the Illinois Department of Healthcare and Family Services, and the Cook County Jail. They built an automated reporting system and a dashboard that staff could use to quickly gain insights into their patient’s needs and quality of care. This automation saved the organization crucial time and resources because collecting and processing this information is often done manually. The new tools allow staff to visualize data in a digestible format, identify new trends, and filter among specific patients to better tailor their care.
They also built out predictive capabilities using patient information from anonymized data streams on nearly 4,000 patients that were treated by Thresholds from 2013–2015. This information included case management files, medical data, and the treatment data for cognitive, wellness, and recovery services. Through these data sources, they were able to develop a system for flagging at-risk patients who need immediate interventions.
Traditional methods of assessing suicide risk have consisted of clinical assessments and questionnaires. These methods have been found to be far less accurate than data-driven models. Researchers from Kaiser Permanent looked at nearly 3 million patients who made mental health and primary care doctor’s visits between 2009–2015. To build a regression model, they pulled data from electronic health records including past suicide attempts, mental health or substance use diagnoses, medical diagnoses, psychiatric medications, inpatient or emergency department care, and depression questionnaires.
The team found that in the 90 days following an office visit, suicide attempts and deaths among patients in the highest 1% of predicted risk were 200 times more common than among those in the bottom half of predicted risk. The study also showed that patients who made doctors visits for mental health issues had risk scores in the top five percent, accounting for 43 percent of suicide attempts and 48 percent of suicide deaths.
Improving Mental Health Care With Data
Research & References of Improving Mental Health Care With Data|A&C Accounting And Tax Services
Source
0 Comments