News Release

Attention to detail can make kids' hospital stays shorter, less costly & less frequent

Quality improvement based on child-specific data yields benefits

Peer-Reviewed Publication

Michigan Medicine - University of Michigan

ANN ARBOR, Mich. -- A new study shows that sick children can go home from the hospital more quickly, less expensively and with less chance of a repeat visit if children's hospitals use a unique source of national data to check their performance against other children's hospitals and show them where they can improve.

The data, from the National Association of Children's Hospitals and Related Institutions, are specially adjusted to reflect the severity, or acuity, of children's illnesses. The study on their use, performed at the University of Michigan's C.S. Mott Children's Hospital, will be published in the October issue of the journal Pediatrics.

The new study is based on data from children who had asthma attacks that required hospitalization. But the authors say their approach could be used to evaluate and improve care of any kind at any children's hospital. If more hospitals paid this kind of attention to detailed data and acted on the results, they could provide higher-quality, more efficient and less costly care to many more children, the researchers say.

The new study shows that the U-M team reduced the length and cost growth of asthmatic children's hospital stays, and their readmission rates, by using child-specific national data to check Mott's performance and look for opportunities to improve the delivery of asthma care in the hospital.

The team used data from the NACHRI Case Mix Comparative Database, which contains information on patients treated at NACHRI member hospitals and adjusts the patients' medical codes by severity of illness and co-existing conditions. This better reflects the nature of children's health problems and the mix of patients, or "case mix" at children's hospitals.

The adjusted medical codes and database allow a children's hospital to make an "apples to apples" comparison of itself against the average performance of all other hospitals. Such comparisons can be made for specific conditions, such as asthma or infectious diseases, and broken down by the severity of the condition. Where there is room for improvement, the data point medical staff toward opportunities for change.

"The original medical data system, called DRGs for Diagnosis-Related Groups, was built for adults, but kids are different," says lead author Aileen Sedman, M.D., who led the team while she was associate chief of clinical affairs at the U-M Health System and who now works as a medical advisor for NACHRI. "Now that we can look at these data in a child-specific way, we have a tremendous opportunity to improve quality."

Sedman and her colleagues started by examining Mott's performance in the NACHRI database. They looked at how Mott's average hospital stay compared with the national average in a broad range of medical conditions. Such comparisons are made possible through the use of all-patient refined diagnosis-related groups, or APR-DRGs, developed by NACHRI in cooperation with the 3M Corporation to reflect children's health problems.

U-M began using APR-DRGs to analyze Mott patient care data in 1998. The next year, Sedman's team in Clinical Affairs found that Mott had an above-average hospital stay for children who had been hospitalized for an asthma attack but had been classed in the least-severe category (level 1) of inpatient asthma patients. The more severely ill asthma patients had shorter-than-average stays.

At Mott, level 1 patients are treated on a general medical service, while more severe cases are treated on a service run by pediatric pulmonologists. Realizing this, the team looked for ways to redesign care on the general service to help children get home sooner and prevent further attacks.

For instance, the pulmonologists developed standardized orders that doctors and nurses could follow for each patient, so that they could adjust levels of inhaled asthma-calming medications throughout the day instead of waiting for a pulmonologist to come.

The team also developed a procedure that automatically notified an asthma educator when a child was admitted to the general floor after an asthma attack. This educational visit helps parents understand how best to manage their child's condition at home and avoid the triggers that can set off an asthma attack and send a child back to the hospital. "We assume that if a child is in the hospital from an asthma attack, something happened that didn't go well and there's a need for more parent education or at-home equipment and medication," says Sedman, an emeritus professor of pediatrics at the U-M Medical School.

Lastly, the team looked in great detail at a sample of medical records, and noticed some consistent problems with the way physicians documented specific information. They worked with physicians to ensure that such information was included correctly, to make sure that the APR-DRG for each child's case was recorded properly.

The study shows Mott's average length of stay in 1999 for level 1 asthma patients was 2.16 days, compared with a national average of 2.14 days. "That's not a huge difference, but since we were far under the national average for more severe patients, we wanted to improve," says Sedman.

After the quality improvement process was in place for three years, the team repeated the comparison. On length of stay, both the national average and the Mott average dropped – but Mott was able to allow children to go home in 1.75 days, compared with a national figure of 2 days. Costs increased on both sides, but Mott contained cost growth to 12 percent, while the growth nationally was 18 percent.

Even more significantly, the education effort seemed to work: the percentage of level 1 asthma patients who were readmitted within 30 days for another asthma attack dropped from 3 percent to less than 1 percent. During the same period, the national readmission rate hovered around 2 percent. Mott also had no deaths among its level 1 asthma patients; there were several nationwide.

"Asthma cases aren't all the same. This process allowed us to ensure that all asthma inpatients, no matter what their severity or what service they're admitted to, are treated appropriately from minute one of their stay, even if they arrive in the middle of the night and aren't seen immediately by a specialist," says co-author Samya Nasr, M.D., clinical director of pediatric pulmonology. "We were able to use the acuity-adjusted data to get the buy-in of all who treat asthma inpatients, and we found that the standard orders and documentation issues created a lot of discussion and interest."

Sedman emphasizes that what works in asthma care could also work for almost any condition that has an APR-DRG code. Already at Mott, the care of certain infectious diseases in children has improved because of changes made after a benchmarking against the NACHRI database. Nasr's team is doing the same for cystic fibrosis patients.

Adult services are also using the acuity-adjusted APR-DRGs -- UMHS chief of clinical affairs Darrell Campbell notes that APR-DRGs adjusted for the U-M's unique patient population are helping guide improvements in care at the main University Hospital. Campbell is senior author on the new paper.

"Before APR-DRGs, data on length and cost of patient stays was useful for hospital administrators who needed to plan and budget, but not appropriate for physicians and nurses who needed to redesign the way they delivered care," says Sedman. "Now, we have a tool that clinicians can use to look at their data after acuity level adjustments have been made, and to use in clinical redesign that's customized to their patients' acuity levels. We just need to encourage its use."

###

Reference: "Clinical Redesign Using All Patient Refined Diagnosis Related Groups" Pediatrics vol. 114 No.4 October 2004.


Disclaimer: AAAS and EurekAlert! are not responsible for the accuracy of news releases posted to EurekAlert! by contributing institutions or for the use of any information through the EurekAlert system.