Why Some Patients Challenge Medical Bills and Others Do Not: What a US Survey Found

What the study measured and why the difference matters

A medical bill can look wrong, be impossible to pay, or simply be hard to understand. Yet a concern on paper does not always become a conversation with the billing office. In a survey analyzed by Erin L. Duffy, Melissa A. Frasco, and Erin Trish in JAMA Health Forum, 203 respondents reported that their household had received a bill they disagreed with or could not afford. Of those respondents, 136 said someone reached out about the concern; 67 said nobody did.

After weighting responses to reflect the adult US population on selected characteristics, the researchers estimated a contact rate of 61.5% and a noncontact rate of 38.5% within that group. Those figures describe what respondents recalled doing about a concerning household bill. They do not show that every contact was a formal billing dispute, that every concern involved an incorrect charge, or that a person who did not contact the office during the survey’s recall period would never seek help.

That distinction changes how to read the findings. The question is not merely whether patients can ask for help. It is who reports seeking help, what the interaction was about, and how much confidence a survey can place in the results people remember receiving.

Who was surveyed and how responses were collected

The researchers drew a random sample of adult participants from the University of Southern California’s Understanding America Study, a probability-based internet panel. Panel members could complete the questionnaire using a computer, tablet, or smartphone. The survey asked about household medical bills received over the preceding 12 months and then followed different paths depending on whether anyone had contacted the billing office.

Of 1,233 panelists invited to this questionnaire, 1,135 completed it, yielding a 92.1% cooperation rate among those invited. This is not the same as saying 92.1% of all US adults responded. The underlying panel had its own recruitment process, and the investigators applied sampling and poststratification weights to better align the survey with selected characteristics of US adults.

The research reported raw respondent counts alongside weighted percentages and, for important estimates, 95% confidence intervals. For example, the estimated 61.5% contact rate had a 95% confidence interval of 51.6% to 70.5%. A count tells us how many people answered a particular question; a weighted percentage estimates a share after accounting for the survey design. They should not be substituted for one another.

The sample was useful for examining a behavior that often disappears between receiving a bill and resolving it. It was less powerful for firm conclusions about narrowly defined subgroups or specific outcomes, where relatively few respondents supplied the relevant answers. Information on characteristics such as financial literacy and personality also came from earlier panel questionnaires rather than measurements taken at the same time as the billing survey.

A bill someone disputes is not necessarily a bill with an error

The first survey question deliberately combined two experiences: a bill the household did not agree with and a bill it could not afford. Later, respondents could select more than one description of the bill that concerned them most. The resulting categories overlap; adding their percentages would count some respondents more than once.

• 46.6% of those reporting a concerning bill said they could not afford to pay it.
• 43.5% felt it was unfairly high.
• 31.8% believed the amount was too high because of a mistake.
• 30.4% said the bill seemed confusing.

These are weighted shares of the 203-person group, not estimates of the percentage of all US medical bills that contain mistakes. Unaffordability concerns the household’s ability to pay; unfairness is an assessment of the charge; confusion is a problem of interpretation. Even a suspicion of a mistake is not the same as a confirmed duplicate charge, coding issue, or payment-posting error.

Consider two possible situations, not cases from the study: one patient may receive an accurately processed but unaffordable bill, while another may see a charge that does not match the services listed in the records. Both could answer the opening survey question in the same way, even though different documents and different responses would ultimately be relevant. Anyone trying to understand the charges and patient balance on a medical bill needs to separate the statement, any insurance explanation of benefits, and the underlying payment or adjustment information. This survey did not perform those checks for participants.

Who contacted the billing office and who did not

Within the group reporting a concerning bill, about three in five said someone contacted the billing office. Contact did not necessarily mean the patient personally made the call. Among those who reported outreach, 104 said they communicated themselves, 35 named another household member, and four said someone from an advocacy organization helped. Because respondents could report more than one person involved, these are not separate, additive groups.

For the 67 respondents who said nobody reached out, the dominant selected explanation was that contacting the office seemed unlikely to change the bill. The weighted estimate was 86.1% in this noncontact group. Other selected reasons included discomfort about contacting the office (34.0%), lack of time (25.7%), concern that speaking up could affect future care (8.7%), and not knowing how to make contact (3.7%). Respondents could choose multiple reasons.

It would be a mistake to turn those answers into a psychological profile. They describe reasons people selected in a questionnaire, not motivations independently established by interviews or experiments. The study also cannot tell whether respondents had already received a satisfactory explanation elsewhere, resolved a different bill, or contacted the office after the recall window.

The statistical models found associations between contacting a billing office and characteristics including education, insurance status, financial literacy, and some personality measures. Respondents without a college degree, people with lower financial literacy, and uninsured respondents were less likely to report outreach after accounting for the variables in the models. These findings highlight unequal reported engagement, but they do not show that any one trait caused someone to call or stay silent. Some group estimates were especially imprecise because the underlying groups were small.

What respondents reported after contacting a provider

Contacting the billing office had several possible purposes. Among respondents reporting outreach, 56.2% said they wanted to ask about price and 47.2% wanted to ask about services. Others sought financial assistance, a payment plan, a lower price, or clarification of an insurance issue. A request for an explanation and a request to change a balance are not the same outcome.

Among the 136 who reported that someone reached out, the study’s weighted estimates for reported results included:

• 25.7% said the bill was corrected (37 respondents).
• 18.2% said they understood the bill better (16 respondents).
• 15.5% reported arranging a payment plan (18 respondents).
• 15.2% reported a price reduction (17 respondents).
• 8.1% reported receiving financial assistance (10 respondents).
• 7.3% said the bill was canceled (six respondents).

These categories could overlap: a person might get an explanation and a payment arrangement, for example. Nor did everyone report a favorable outcome. The survey also recorded 32 respondents (21.8% weighted) who said the problem had not yet been solved and 35 (23.9% weighted) who reported that nothing changed. Those answers should not be added to the other categories as though the responses were mutually exclusive.

The researchers also looked at smaller groups defined by the reason for contact and whether the issue was reported as resolved. Within one such subgroup, 25 of 37 respondents who suspected a billing mistake said it was corrected, corresponding to a weighted estimate of 73.7%. Another analysis reported financial relief among 75.8% of an affordability-related subgroup. Those narrower figures are easy to misread. They do not mean that 73.7% of all suspected billing errors in the country are corrected, or that 75.8% of anyone who calls about an unaffordable bill will receive relief. The subgroup selection, small counts, and self-reporting all matter.

The study did not inspect a corrected invoice, verify the patient’s new balance, or establish that the action taken by the billing office alone produced the reported change. A payment plan can make a balance more manageable without lowering it. A corrected bill can reflect a legitimate adjustment, but the survey cannot independently tell which accounting event occurred.

Why the findings do not prove a universal dispute success rate

The safest way to interpret these figures is to keep what was measured separate from what is inferred. A survey can describe reported actions and experiences with reasonable care while still leaving important questions unanswered.

Two pieces of wording show why interpretation requires care. The opening question included bills people “did not agree with or could not afford to pay.” That is a broad measure of financial or billing concern, not an error audit. Among people who did not contact the office, many selected the reason “did not think it would change the bill.” That records a stated expectation; it does not prove the expectation was correct, irrational, or the only reason they chose not to contact anyone.

There is a further selection problem in comparing people who called with people who did not. People with different types of bills, time constraints, financial resources, or prior experience might choose differently. The reported outcomes among callers therefore cannot be treated as the results that the noncallers would have achieved had they called. The researchers used statistical adjustments to examine associations, not a randomized experiment to identify the causal effect of contact.

Reported percentages also have uncertainty. The 95% confidence interval around the overall contact estimate was wide enough to matter, and the intervals around small outcome subgroups were wider still. A headline that removes the denominator and these limits can turn an interesting observation into an unsupported promise.

What stronger research could test next

A stronger study of billing-office outreach would first define the outcome precisely. Was the goal an itemized explanation, a verified correction, a discount, an assistance award, or a workable payment schedule? Those are different events, and combining them under the word success hides the question that matters.

Researchers could then pair a larger and more varied patient sample with billing records, insurer explanations of benefits where applicable, and documented account changes. With appropriate consent and privacy safeguards, the study could distinguish a charge that was found to be wrong from a valid charge that was reduced for financial reasons. It could also track the time spent, the number of contacts required, and whether a case remained open.

Another approach would be to test support offered before patients need to initiate a call. For example, a carefully designed trial could compare a standard statement with a statement that clearly identifies the billing contact, explains how to request an itemized bill, and presents available financial-assistance information. Researchers would need to measure verified follow-through and outcomes, not just whether recipients found the statement clearer. Such a test could evaluate whether better information reduces administrative friction; the existing survey does not show that it does.

Follow-up over a longer period would also address a limitation of one-time recall. A bill may be corrected after several conversations, or a person may contact the office only after reviewing records or receiving another statement. A single survey answer may capture an unfinished process.

Final thoughts: why access to billing help differs

This survey found a meaningful difference between reporting a troubling medical bill and reporting contact with the office that issued it. It also documented different reported outcomes and uneven patterns of outreach across respondent groups. That is evidence about patient behavior and administrative burden, not a verified tally of incorrect bills or proof that calling guarantees relief.

For researchers and policymakers, the next question is how much work patients must do to understand a charge, find the right contact, and obtain a documented answer. For someone dealing with a specific provider charge, the appropriate next resource is a practical guide to checking and disputing a medical bill with the billing office. The survey can explain why that process deserves attention; it cannot decide the outcome of an individual account.

Frequently asked questions