By Yanqi Xu

Hospitals should deliver quality health care to patients just like smartphones to their users. That’s the analogy Dr. Anupam Jena used to open this session.

The care should be fast and effective, so patients receive care as soon as possible. It should be personalizable, too, because everyone’s willingness to take risks is different. We want our smartphones to be efficient, Dr. Jena said. in the health care context, we want the care to be affordable, so we don’t spend an astronomical amount on something that shouldn’t be that expensive. Last but not least, we strive to make our technology and our health care equitable, so they become available across racial and socio-demographics. 

These factors underlie quality health care, which researchers like Jena have been trying to measure for years. 

However, due to a lack of random patient samples, it’s not easy to directly measure the structure, processes and outcomes of care and reach conclusions about quality. Does a hospital have higher mortality because it treats higher-risk patients? Or are there other factors?

Using the traditional scientific method to randomly assign patients to a certain hospital is problematic, Jena said. But academics have identified certain natural experiments in real life: For example, researchers have studied patients taken to hospitals by ambulances, as which ambulance company gets dispatched to pick up patients can be random. Another case worth studying is days with road closures like during marathons when patients would be more likely taken to the closest hospital. 

Moderator Melanie Evans noted that measuring health care quality is tied to hospital rankings and ratings, for example, by the Centers for Medicare & Medicaid Services, which many consumers rely on. It can get wonky, though, so journalists rely on academics to help them evaluate these resources and data.

Annie Waldman showed how she used data to assess maternal health mortality. Her 2018 project Lost Mothers, which focused on maternal mortality disparities, went beyond the anecdotes to look at the true scope and extent of maternal harm. She used hospital discharge data in some states and decided to look at one cause of preventable deaths — hemorrhage. 

Hospitals typically treat hemorrhage with a special kit that should be easily accessible. Using the coding of hemorrhage in the data as a proxy, ProPublica examined how the larger health care system failed to protect patients and how Black people had the highest mortality rate. The newsroom used an algorithm already developed by the Alliance for Innovation on Maternal Health and paired rigorous data analysis with interviews with grieving family members.

She found personal stories from social media, and Yelp, Facebook and Google reviews, related hashtags such as #RIP or #mom, malpractice lawsuits, closed Facebook groups, as well as GoFundMe pages.

Ariana Tobin worked closely with Waldman on the Lost Mothers series as the crowdsourcing and engagement team editor. 

She and the engagement team designed the social callout on their website and brainstormed strategies for reaching an audience that has been historically underrepresented. They also helped sources file records requests to gather receipts, medical reports, photos and other evidence under strict HIPAA rules.  

In some other cases, the engagement team did community research and pre-reporting, trust building, gut checking, and user testing. In a project about medical debt in Memphis, reporters wrote hand-written letters to patients who might be afraid to pick up their phones when an unknown number called. 

Tobin said it’s very important to keep in touch with these sources. A big source list is an invaluable resource for journalists to keep developing story ideas and reporting on marginalized communities. 

Yanqi Xu is a reporter at the Flatwater Free Press. She was a 2023 AHCJ-Rural Health Journalism Fellow.