April 2001 Bulletin

Geographic variation linked to patients

Study finds individuals in high-rate area more willing to have surgery

A new study on geographic variations in surgery has found that the patient plays a major role in this confusing issue.

Other studies have looked at many aspects of geographic variation, but researchers say this is the first to examine the patient’s willingness to have surgery.

Variation in surgery rates by census region or even nearby communities has both health policy and public relations issues. Without a plausible explanation, it is usually assumed that high rates of surgery indicate over-utilization of resources by aggressive surgeons.

It also is assumed by the media and the public that the right rate of surgery is the low rate, or perhaps the right rate is somewhere between the low and high rates.

One study of 35 million Medicare patients suggests that whether or not you have orthopaedic surgery depends a great deal on who you are and where you live.

Richard H. Gelberman, MD, AAOS president, pointed out at the Annual Meeting last month that a patient in one city in the central valley of California is 50 percent more likely to have a spine operation than a patient in San Francisco, less than 100 miles away. Patients in the central valley city are four to five times more likely to have a total joint replacement than patients in San Francisco. "Why?" he asked.

Another study in Canada suggests there may be unmet needs in both the low-rate and high-rate communities. "If that’s the case, we need to mount a public health initiative to inform primary care providers of the benefits of our care," said Dr. Gelberman.

Still another study has found that the dominant variant in rates of surgery is the orthopaedic surgeon’s enthusiasm for the procedure.

Dr. Gelberman told the members that "we must gather the evidence and then address this issue."

The new study by Canadian researchers puts, for the first time, the patient’s willingness to have hip and knee arthroplasty in the equation.

The researchers estimated the need for and willingness to have arthroplasty in high- and low-utilization areas of Ontario, Canada. They documented a wide gap between potential need as defined from the point of view of surgical indications and actual need as defined from the point of view of the patient.

Among those with severe arthritis, only 8.5 percent in the low-rate area and 14.9 percent in the high-rate area were willing to have joint arthroplasty.

The researchers didn’t just collect statistics from a databank and X-rays from the repository. They went to participants in the study to examine and interview them, said study coauthor, James G. Wright, MD, MPH, the R.B. Salter chair in surgical research and a professor, faculty of medicine, University of Toronto.

The researchers focused on residents of Oxford county, which ranks among the top six for rates of hip and knee arthroplasty among the 49 Ontario counties, and residents of East York, which ranks among the bottom three for arthroplasty rates. The rural farming communities are within three hours driving time from Toronto.

The researchers determined arthritis severity and comorbidity by questionnaire, established the presence of arthritis by examination and radiographs and established the participants’ willingness to have arthroplasty in interviews.

They found demonstrable need and willingness were greater in the high-rate area, suggesting these factors explain, in part, the observed geographic variations for this procedure, according to the study in the March issue of Medical Care.

There were 19,675 individuals eligible for the study in the low-rate area and 19,678 in the high rate area. After winnowing the eligible candidates to meet their criteria, the researchers found 401 of 447 individuals in the low-rate area and 520 of 580 in the high-rate area were considered to have potential need for arthroplasty. Expressed per 1,000 of the eligible individuals (19,675 and 19,678, respectively), the potential need for arthroplasty was 28.5 in the low-rate area, compared with 36.2 in the high-rate area.

Among those with severe arthritis who were not on a waiting list for arthroplasty, 8.5 percent in the low-rate area and 14.9 percent in the high-rate area were "definitely willing" to have arthroplasty. Of the remaining, 17.5 percent and 21.5 percent, respectively, in the low- and high-rate areas were either "probably" or "definitely" unwilling to have arthroplasty.

Estimates of need for arthroplasty, adjusted for willingness and expressed per 1,000 of the eligible individuals (19,675 and 19,678) were 2.4 in the low-rate area, and 5.4 in the high rate area.

Those in the high rate area were more likely to have known someone who had undergone joint arthroplasty than those in the low-rate area. However, both groups believed the individuals who had arthroplasty were satisfied with the results.

The differences in willingness to have the surgery may reflect a perception that arthritis is just part of aging and should be accepted, said Dr. Wright. Some individuals may not consider the surgery because they fear they will become housebound or because they don’t have a social support structure, he observed.

"The most important predictor of definitive willingness to have arthroplasty was having previously discussed this procedure with a physician, emphasizing the importance of patient-physician interaction in patient decision-making regarding surgery," the researchers said.

In a 1999 study on determinants of regional variation in the use of knee replacement, Dr. Wright and colleagues concluded physician enthusiasm was the controlling factor. Their conclusion was "after controlling for population characteristics and access to care (including the number of hospital beds, and the density of orthopaedic and referring physicians), orthopaedic surgeons’ opinions or enthusiasm for the procedure was the dominant modifiable determinant of area variation."

Dr. Wright says that conclusion was based on a belief that there was no variation in disease prevalence. Further research found there is variation of disease by area.

The two studies are not in conflict, he said, because both the patient’s willingness and the interaction with the physician are factors.

Estimated potential need for and willingness to have hip/knee arthroplasty*

Low-Rate Area


Rate per 1,000


Rate per 1,000


potential need for arthroplasty†

definitely willing to have arthroplasty

willing to have arthroplasty‡

55—64 years




65—74 years




75+ years








High-Rate Area


Rate per 1,000


Rate per 1,000


potential need for arthroplasty

definitely willing to have arthroplasty

willing to have arthroplasty

55—64 years




65—74 years




75+ years








* Represents approximate estimates of potential need for arthroplasty per 1, 000 population

† Rate per 1,000 respondents, potential need for arthroplasty defined as severe arthritis (WOMAC 39/100), evidence of arthritis of the hip or knee on radiographs and clinical examination, and no absolute contraindication to having arthroplasty.

‡ Rate per 1,000 respondents, willingness to have arthroplasty defined as ‘definitely’ willing to have the procedure based on Phase III interview.

Study on patient preferences

In Phase I of the study, a questionnaire asked participants to report the presence of symptomatic joints on a diagram (homunculus), the presence or absence of specific functional disabilities, and whether they had undergone prior arthroplasty. Respondents were selected for Phase II if they had at least moderately severe hip or knee complaints as defined by difficulty in the last three months with each of stair climbing, arising from a chair, standing and walking; swelling, pain or stiffness in any joint lasting at least six weeks in the past three months; and indication on the diagram that a hip and/or knee was "troublesome."

The Phase II respondents completed three measures of the severity of their hip and knee complaints: the Western Ontario McMaster University Osteoarthritis Index (WOMAC), the Health Survey Short-Form (SF-36), and the disability subscale of the Health Assessment Questionnaire (HAQ). They also reported (from a list of 18) concurrent health problems for which they were receiving treatment or had seen a physician in the past year.

In Phase III, the researchers visited all East York respondents with WOMAC scores 39/100 to determine the positive predictive value of the WOMAC summary score cutoff for identifying individuals with "arthritis." They assumed the proportion of patients with X-ray and clinically confirmed hip and knee arthritis would be the same for both regions. The East York participants underwent standardized joint examination in their homes.

Each hip and knee was classified as "arthritic" if there was stress pain, reduced range of motion, or deformity on examination and any of osteophytes, joint space narrowing, subchondral sclerosis, marginal erosions or subchondral cysts on radiographs.

A standardized interview, based on audiotapes of 10 patient-surgeon discussions, was designed to emulate the typical conversation between a patient and surgeon. The interview described the consequences of not having surgery, alternative treatments to arthroplasty, and the risks and benefits of arthroplasty, including the projected life span of the replaced joint.

Potential risks associated with arthroplasty were discussed in greater detail than usually provided by orthopaedic surgeons in order to provide a conservative estimate of willingness to have arthroplasty. Participants indicated their willingness to have arthroplasty on a five-point scale: "definitely not willing", "probably not willing", "unsure", "probably willing", and "definitely willing". Joint examinations and interviews were performed by trained Arthritis Society physiotherapists.

Home Previous Page