April 1997 € Volume 24 € Number 2

R_x for U.K. Healthcare Woes

This is another in a series of articles based on interviews with recent Franz Edelman Award finalists. Geared toward practitioners, the articles strive to provide lessons the Edelman authors learned and some pitfalls they encountered during the course of their work.

York University team develops new formula for distributing $35 billion in funds

By Nancy Bistritz

Healthcare, and a proposed national healthcare system, have been a topic of hot debate in the United States for several years. The question of how much control the government should have when it comes to an individual's healthcare needs has caused tension among politicians and insecurities among those who would have to play by the system's rules. Healthcare has also been a source of much debate in the United Kingdom, where a national system has been in place for many years. The recent controversy in the U.K., however, has not been about who controls the system, but rather how the system allocates its funds.

In 1948, the U.K. introduced the National Health Service, an organization -- funded primarily out of general taxation -- which allocates funds to 105 different local health authorities with average populations of 500,000. Each year, the NHS Executive (the organization which manages the NHS and is answerable to the Secretary of State for Health) sets a "cash-limited" hospital and community health service (HCHS) budget, which, in turn, is distributed to the local health authorities. Using these funds, the local health authorities finance the hospital and community health care as needed by the patients and as deemed necessary by the family practitioners.

Patients in the United Kingdom may only seek care in an NHS hospital if they are referred there by their family practitioner. Additionally, with only a limited amount of funds in each authority, those patients referred to hospitals (with the exception of emergency care) are not always treated immediately and are, in some cases, placed on a waiting list.

Since its inception, the method by which the NHS allocates funds has come under fire. The thorny question: What is the most fair and equitable way to distribute $35.2 billion annually to 100 local authorities with different healthcare needs and different populations?

In 1970, in an effort to answer this question, a formula was implemented based on population, the differences in the need for healthcare, and the geographical differences in the cost of providing healthcare services [Department of Health and Social Security 1976].Due to variations which existed in the use of NHS funds by different age/sex groups, the population category was separated again by age and sex and then was weighted by an index. At the time, experts saw no other useful index besides morbidity and recommended using the specific standardized mortality ratio (SMR) as a determinant for needs.

Over the years, one of the biggest criticisms of the formula was the use of mortality rates as an indicator of healthcare needs since, at the time, there was no empirical evidence to indicate that SMRs related to health care needs.

In 1985 another review was implemented, this time using social variables. In 1993, a team of professors from York University was asked to specifically examine the SMR as it relates to the need index. Peter Smith, professor of economics at York University, Treavor Sheldon, Roy Carr-Hill, Geoffrey Hardman, Stephen Martin and Stuart Peacock began collecting more recent and complete data.

Methods Used
"One of the main functions of the York University work was to try to bring in social deprivation which has an effect on demand for health services and need above morbidity," said Nick York, economics advisor in charge of resource allocation at the NHS Executive.

In order to do this, the York University team began collecting data on every hospital admittance in the United Kingdom over a three-year period. Specifically, the team was looking at the location in which the admittance occurred, and the individual who was admitted. Consultations with experts and the NHS revealed that small, individual studies, were likely to be relevant. Therefore the team decided to study small areas with populations of about 10,000. Using these units of analysis, data were collected relating to each area's socioeconomic conditions, the existence of health services and the use of inpatient services [British Medical Journal, pp. 1047].

They split the healthcare system into several categories, specifically looking at the acute care sector and the psychiatric care sector. The team found it could adequately model the use of healthcare using the following variables for the acute care sector: mortality rates within a given area, long-standing illness, the proportion of the elderly who were homebound, single-parent families, and unemployment.

According to Smith, those variables explained well over 50 percent of all healthcare utilization in the country.

"We found the variables by using multi-level techniques. We could combine them into an index of healthcare needs so that the index demonstrated the relative healthcare needs of every area in the country," Smith said. "The intention was to try and link the cost of healthcare from these small areas to the social conditions in those areas. We then looked at various social variables -- about 60 -- such as no car ownership, long-term illness, single-parent families, etc. We looked to see which of those variables were the most heavily correlated with healthcare utilization, with the intention that those that were the most closely correlated were going to be the ones that would form the basis of our formula for distributing healthcare funding."

For the psychiatric sector, the York University team found there were six social variables that contributed to the utilization of healthcare:mortality, single parent families, ethnicity, proportion of dependent persons with no caregiver (typically the elderly with limiting illnesses who had no one to care for them), the elderly living alone, and the population deemed permanently sick.

The variables identified were for both acute and psychiatric inpatient care, but did not answer questions about noninpatient care. Research indicated that using the same indices for noninpatient services as inpatient services could result in heavy cash losses and/or gains -- as much as 5 percent. However, in the absence of a better method, the York University team recommended that the NHS use the formula it had developed for the acute sector and then commission new research to look at those additional services. The NHS, however, decided to take a different route.

"It decided not to use a needs formula at all for the 24 percent remaining NHS expenditure," Smith said. "This caused quite a deal of controversy because it was really saying that every area in the country, per capita, had the same need for community health services. I think a lot of commentators thought that didn't sound right because there were clearly areas where there appeared to be more need for community services than others."

After a review by the House of Commons -- at which the York University team presented evidence -- the Secretary of State for Health agreed that needs weighting would be used as part of the formula for distributing funds in several areas that had not used this basis before. "We feel we have persuaded policy makers of the desirability of using these techniques throughout the entire National Health Service," Smith said.

Major Hurdles
While the preliminary work took only four months, the refinement would last another 14. Along the way, the York University team faced several obstacles that challenged them even further.

"One of the major problems we had was that there was a widespread belief that it wasn't just healthcare needs that influenced utilization, but also the supply of hospitals and other healthcare," Smith said. "One might argue that those who live near a hospital are more likely to use it, and one of our biggest challenges was how to adjust for that. That occupied a lot of our time during this study."

Still another challenge Smith and his team faced was the possibility of bias in the data. Because the team was using smaller areas as units of analysis, the concern was that the policies of the controlling health authorities might affect the small areas within their boundaries. "A health authority might have 50 small areas, and we were concerned that because of that, the data coming from those 50 areas might be systematically biased for whatever reason because they lay in that particular health authority," Smith said.

Because of this, the team used multi-level modeling or hierarchical modeling -- a fairly new technique in the policy-making arena. "They (policy makers) were used to simple regression analysis, and we had to spend a great deal of energy trying to persuade the policy makers that multi-level modeling was the correct approach," Smith said.Using multi-level techniques, the team was able to combine the variables into an index of healthcare needs in order to demonstrate the relative healthcare needs of every area in the country.

Throughout the 18-month period, the York University team had to remember it was dealing with an issue that had, at times, been discussed, debated and criticized. Additionally, it needed to keep in mind that the issue of healthcare was not only a sensitive one, but a political one as well. However, the response they received while working on such a delicate issue was anything but unsupportive. "I think our biggest surprise was how receptive people in the NHS were to the results we presented because the nature of what we were doing was that we weren't increasing the pot of money at all, we were just redistributing it," Smith said. "So for every winner, there was going to be a loser, and that's quite a difficult situation to be in."

He attributed the reaction of the NHS to a new-found acceptance to the field of management science. "I think in some ways that (response) demonstrates that management science can be a very powerful influence if it's really persuasive, because it becomes very difficult for the losers to argue convincingly against your proposal."

Lessons Learned
Throughout the 18-month long process, the York University team fine-tuned the formula several times going back to the NHS time and time again with new data and new findings. Dealing with an issue that had, at times, been controversial and had the potential to stir up mixed opinions, the York University team found itself under extreme pressure.

"There was a lot of pressure on us throughout the study -- each time we developed a different model," Smith said. "There was a lot of pressure for us to indicate what the financial implications would be for different areas of that new model. I think it was very helpful for us not to get involved (with the financial implications) until we felt we had developed a satisfactory model." Under the advice of one of the team's external advisors, all of the technical work was completed and accepted by the client prior to looking at what the policy implications would be.

Since the formula and its indices were presented to the NHS, there has been some additional work done, but not much. As Nick York explained, there has been much "tweaking" of the formula, but most of the technical work is complete, leaving little room for adjustments. "The work that the York University team carried out has had quite an impact over the last two or three years over the distribution of funds -- mainly moving money toward the inner cities," York said

He does admit that because of the comprehensive work done at York University, there is little need for adjustments.
Nancy Bistritz is the assistant editor of OR/MS Today
For more information, put the number 21 in the appropriate space on the
Reader Service Form

E-mail to the Editorial Department of OR/MS Today: orms@lionhrtpub.com

Lionheart Publishing, Inc.
2555 Cumberland Parkway, Suite 299, Atlanta, GA 30339 USA
Phone: 770-431-0867 | Fax: 770-432-6969
E-mail: lpi@lionhrtpub.com

Rx for U.K. Healthcare Woes

By Nancy Bistritz

R_x for U.K. Healthcare Woes