Health scoring: Can scoring improve patient care and survival?

• Author:
• Author name

  Quantumrun Foresight
• October 7, 2022

Insight summary

Health scoring, a tool for assessing patient risk, is reshaping healthcare by enabling more targeted treatments but faces challenges in ensuring fairness across diverse populations. These scores, calculated from data like electronic health records, are increasingly used to predict patient outcomes and manage diseases, yet their reliance on specific demographic data can limit their universal applicability. As artificial intelligence (AI) advances, it offers potential improvements in scoring accuracy and speed but also raises concerns about privacy and discrimination.

Health scoring context

The COVID-19 pandemic has shown the importance of accurately screening patients to ensure correct diagnoses and timely treatments. Health scoring has been proven to identify many patients’ prognoses correctly. As artificial intelligence (AI) systems continue to automate numerous areas in healthcare, they might soon be able to provide more accurate health scoring systems.

Health scores, also known as risk scores, classify people for targeted healthcare testing and treatment. These assessments compute an individual’s score based on risk factors data; a higher score indicates greater danger. Health scores are calculated by software that analyzes routinely stored data (electronic health records) and may be used to evaluate individuals or populations. The information can help primary care facilities send screening invitations to those at risk of developing a specific condition. Some chronic diseases have established health scores that allow physicians to create a treatment plan based on how the disease is anticipated to progress.

However, one major challenge in developing health scores is that they tend to have limited or one-sided methodologies. For example, the UK’s Cambridge Diabetes Risk Score aims to detect undiagnosed type 2 diabetes mellitus (T2DM). It collects data on age, sex, body mass index (BMI), steroid and antihypertensive medication use, family history, and smoking status.

However, while this score would be suitable for use in primary care, it does not reflect the higher incidence of T2DM in those from black and minority ethnic (BME) groups. Similarly, in 2003 the Finnish Diabetes Risk Score (FINDRISC) was established. The FINDRISC heavily leaned into the European population and was unsuitable for use in diverse communities. Thus, health scores need to be developed using diverse populations to ensure they can be applied universally.

Disruptive impact

Between 2020 and 2022, scoring systems aided healthcare facilities in predicting which COVID-19 patients would require mechanical ventilation. The score was comprised of three components: heart rate, the proportion of oxygen saturation, and a positive troponin I level. The first two are readily available from vital signs, and the third was often obtained through routine lab tests, allowing this assessment method to be done at any hospital. This scoring system greatly assisted healthcare workers in managing cases as hospitals grew overburdened.

In October 2020, the Massachusetts General Hospital used artificial intelligence (AI) to create a multi-factor score (COVID-19 Acuity Score (CoVA)) to forecast the prognosis of COVID-19 patients in urgent care or emergency departments. The score assesses how likely the patient is to develop complications or require hospitalization. The top-five factors that determined the result were age, diastolic blood pressure, blood oxygen saturation, COVID-19 testing status, and respiratory rate.

In general, healthcare professionals included 30 variables in the scoring model, including medical history, chest X-ray findings (when available), and heart function. According to the researchers, CoVA was developed to enable electronic medical record systems to use automated scoring. The score was helpful during COVID-19 outbreaks when quick clinical assessments make a lifesaving difference.

Implications of health scoring

Wider implications of health scoring may include: 

• Healthcare providers using scoring to predict prognoses for diseases like diabetes and heart illnesses.
• Community resistance against its use. Patients and community members may believe that health scoring is used to heighten discrimination in healthcare. 
• More hospitals and healthcare networks partnering with AI companies to develop accurate models for scoring.
• AI systems being used to further develop health scores that can help prioritize treatments and reduce diagnosis delays.
• Hospitals and research centers collaborating globally to come up with standards on health scores for future pandemics and epidemics.
• Healthcare insurance companies adjusting premiums based on health scores, leading to more personalized pricing but also potential concerns over fairness and privacy.
• Governments reevaluating public health strategies to incorporate health scores, improving population health management.
• Public skepticism growing around health scoring, leading to new regulations on how health data is used and shared.

Questions to consider

• How can hospitals ensure that health scores are accurate and ethical?
• What are the other challenges in implementing health scores for disease assessment?