Obesity and Smoking, Alcohol Use Underreported in National Database

The Nationwide Inpatient Sample (NIS) has been critically underreporting certain health information for its database, according to a recent study by Johns Hopkins researchers. The misreports, mostly in patient obesity status and alcohol/tobacco use, may be incentivized by monetary reasons.

NIS data, consisting of billing data from 20% of patients discharged from randomly selected hospitals across the country, is used by CMS to make national estimates of healthcare utilization, access, charges, quality, and outcomes. Hospitals who are determined to have a higher risk for readmission or surgical complications usually receive lower reimbursements. Obesity status and alcohol/tobacco use, 2 factors that are related to higher risks, are often not recorded in the database.

“Hospitals that are better at using billing codes to record the alcohol, tobacco, and obesity status of their patients will be classified as having higher-risk patients,” Susan Hutfless, PhD, MS, of the Johns Hopkins University School of Medicine, said in a statement. “Their reimbursement will be lower than hospitals that don’t include the information in their billing records.”

The level of underreporting in the NIS is seen most dramatically when compared with other databases, such as the Behavioral Risk Factor Surveillance System (BRFSS). For instance, obesity prevalence in the US is recorded as 27.4% in the BRFSS, but as only 9.6% in the NIS. The BRFSS records 20.1% of the country as current smokers and 18.3% as heavy drinkers, while the NIS records only 12.2% and 4.6%, respectively. In all methods of data comparison (national vs state-by-state), NIS statistics were substantially lower than BRFSS’. (BRFSS data is extracted from a phone survey of over 500,000 Americans about their health, whereas NIS data is taken from patients’ hospital bills.)

This discrepancy in information could not only lead to a negative economic impact on hospitals who record correct information and actually need the most reimbursement, but can also lead to inaccurate national health quality assessments.

“If there are problems with coding obesity and tobacco, which are 2 two greatest contributors to mortality worldwide, it is reasonable to assume that there may be errors in the way we perceive the burdens of certain diseases on the US healthcare system,” Dr Hutfless said.

And the fact that this discrepancy shows itself in 2 administrative databases that are both considered to be nationally representative is means for concern, according to the authors. The study, the first to estimate the extent to which administrative databases may be under-coding important health indicators, is a step towards a potential solution.

“Improving the accuracy and the utility of information in administrative databases like NIS will contribute to our ability to affect health care decisions or health policy decisions that are heavily based on the findings from these sources,” authors wrote.