Tradeoffs collected and analyzed financial assistance policies of every nonprofit general hospital and children’s hospital in Texas to evaluate how they proactively assess need and award financial aid. We excluded hospitals that specialize in psychiatric care, rehabilitation and long-term acute care.
Tradeoffs identified 166 nonprofit facilities in Texas using state records and federal Centers for Medicare and Medicaid Services data as of Sept. 29, 2026.
To find the policies of all nonprofit hospitals in Texas, Tradeoffs started with a list of hospital financial assistance policy websites maintained by the nonprofit Dollar For. To verify websites were up to date and fill in missing policies, Tradeoffs downloaded policies from hospital websites and contacted hospitals directly. When hospitals did not share their policies, Tradeoffs filed public records requests to the state of Texas, which requires hospitals to submit policies annually.
Tradeoffs then used artificial intelligence and data analysis tools to extract and organize the contents of hundreds of pages of hospital financial assistance policies. The use of AI allowed for an efficient preliminary analysis and the creation of organizational tools to enable human review.
To analyze the policies, we used a large language model to review the policies for yes/no answers to the following questions. In every case, we asked the AI to include the exact text snippet serving as evidence.
- Does the document contain an organization name?
- Does the policy include an effective date or a date for its most recent revision, approval, or update?
- Does the policy allow the use of presumptive eligibility to grant financial assistance or charity care?
- Does the policy automatically grant financial assistance/charity for those screened and deemed presumptively eligible without an application or documentation from the patient?
- Does the policy require presumptive eligibility screening before the hospital ever sends a bill?
- Does the policy require presumptive eligibility screening after sending bills, but before the hospital takes collection action for medical debt?
- Does the policy require presumptive eligibility after sending bills to collection, but before writing off the bill as uncollectible?
- Does the policy law set any requirement for when presumptive eligibility screening must occur?
- Must hospitals notify patients of presumptive eligibility results?
- Does the policy require a refund to patient were awarded financial assistance/charity but who had already paid some or all of their bill?
Finally, Tradeoffs asked AI to identify any criteria for patients eligible for presumptive eligibility contained in the policies. We identified the following groups of criteria:
Financial resources: Other income-based aid programs; low-income housing; employment status; vow of poverty; bankruptcy; patients say they can’t pay; prior approval for financial assistance; financial assessment by third-party vendor.
Health: Deceased (with exceptions); mentally or physically incapacitated.
Housing status: No housing; incarcerated.
Insurance status: Medicaid; self-pay; uninsured; other insurance source (including low-income Medicare recipient or exhausted commercial benefits).
Uncollectable or unreachable: Returned by collection agency; history of non-payment; can’t be reached; failed attempts to collect application.
Other: Crime victims; at the discretion of hospital staff; citizenship status.
The AI produced an output with citations of verbatim text from policies. We then used a text-matching program to digitally verify the text. Finally, Tradeoffs journalists manually reviewed and either confirmed or corrected the results.
Tradeoffs journalists analyzed the data and interviewed accounting and hospital billing experts to understand the significance and timing of presumptive eligibility choices.
Tradeoffs then emailed each hospital or health system our findings for their financial assistance policies and asked them to confirm or clarify the results. We followed up by email when hospital responses required further clarification.
We made repeated attempts by email and phone to contact hospitals that didn’t respond to fact-checking requests. Two Tradeoffs journalists independently reviewed the results for every hospital that didn’t respond to fact-checking.