Charts released by California state health planners Monday show enormous frequency variation in four surgical procedures in each of 58 counties, a finding sure to raise questions about regional appropriateness of care throughout the state.
The charts were released for each of three years, 2005, 2006, and 2007, by the Office of Statewide Health Planning and Development (OSHPD), the agency that collects and researches discharge data for nearly every hospital in the state. The statistical collection is said to be the largest state quality database in the country.
Among the 13 counties with the state's largest populations—more than 800,000—for example, the charts received some wide and unexplained differences. Hospitals in Kern County performed the most percutaneous coronary intervention or stent procedures in 2007, 512 per 100,000 admissions, which was 46% more than the state average and nearly double the number performed in Contra Costa County. San Diego County was second with 437.1 stents per 100,000 admissions.
Fresno County performed the most coronary artery bypass graft procedures in 2007, 186.6 per 100,000 admissions—50% more than the state average and 254% more than San Francisco.
Kern County also performed more hysterectomies than the 13 other large counties, 544.2 procedures per 100,000 admissions, while Ventura County performed the most laminectomy or spinal fusion procedures, 336.2 per 100,000 admissions. State averages for those procedures in 2007 were 321.3 and 201.1 respectively. Among the most populated counties, San Francisco County performed the fewest hysterectomies and the fewest laminectomies, 149.4 and 126 per 100,000 admissions.
The state charts are an effort to publish information about regional utilization similar to reports produced by the Dartmouth Atlas, which looks at regional variations in medical practice and healthcare utilization and cost, especially in the last months of life.
But while the data is risk-adjusted for age and sex, it is not corrected for severity of disease or co-morbidities, defects that make the data hard to interpret, acknowledges Joe Parker, director of OSHPD's Healthcare Outcomes Center. "There's a lot of unexplained variation and the numbers change quite a bit," Parker says.
He adds even though these are "fairly crude numbers" it may indicate that there are doctors practicing medicine in ways "that are very divergent . . . raising interesting and useful questions that policymakers, or county public health officials or anyone else might like to ask."
For example, it is unclear from the data whether counties where practitioners perform the lowest numbers of these four procedures, such as Alameda, Los Angeles or San Francisco, put up barriers to access, or whether their residents are just healthier and need such procedures less frequently. Or, perhaps, in those counties, standards of practice or less aggressive doctors and hospital policies hold back physicians from performing the procedures on patients for whom any benefit is less probable.
Likewise it is unclear whether providers in Kern, Fresno, and Riverside counties, which are above the state average in all four procedures, are overutilizing care, or whether their patients are just a lot sicker. Or, perhaps, those counties may have fewer barriers to care for the poor.
"This might be a starting point for researchers who want to delve into this deeper," says Debby Rogers, RN, vice president for the California Hospital Association. But she adds that it isn't very useful to hospitals because it is not hospital specific.
Parker indicates one of the surgical categories examined, laminectomy or spinal fusion, does offer a clue about regional variation in practices. "For laminectomy or spinal fusion, the evidence is not as robust that people get a whole lot of benefit." He adds he would be "more suspicious" if he saw a lot of those procedures.
The same goes for hysterectomies. Research suggests that hysterectomies, one of the most common and controversial surgical procedures in the country, is often unnecessarily performed to correct endometriosis, which can usually be corrected through easier methods, Parker suggests.
Brian Paciotti, an OSHPD research scientist who specializes in utilization statistics, emphasizes that his agency is not pointing fingers at any hospital system or provider group for anomalies in care. But he says for those counties where there are large differences, "it's up to the people, and the stakeholders, in these counties to drill down and take a look at why."
The state also released three other county-by-county charts listing:
- Pediatric quality indicators, including rates of short-term diabetes complications, asthma, gastroenteritis, perforated appendix and urinary tract infections
- Prevention indicators, including rates of uncontrolled diabetes, bacterial pneumonia, congestive heart failure, hypertension, and diabetic lower extremity amputations
- Patient safety indicators, indicating avoidable hospital events, such as postoperative hematoma or hemorrhage, postoperative wound problems, infections related to care, and pneumothorax related to care.
The 2007 charts can be compared with 2006 or 2005 to see if any of practice patterns may be shifting.