[Skip to content]

Sign up for our daily newsletter
The Actuary The magazine of the Institute & Faculty of Actuaries

Chronic disease management

Disease management (DM) aims to improve the overall health of the chronically ill, as measured by better clinical outcomes, cost savings, and improvements in the quality of life of the patient.
DM is considered to have two components:
– helping patients with chronic illness to self-manage the disease, through education and communication; and
– encouraging clinicians to improve their care of such patients through education, better and more timely information, and by using evidence-based medical principles.

Principles of disease management
The chronic illnesses selected for DM programmes commonly include asthma, diabetes, congestive heart failure, and coronary artery disease, as well as other respiratory diseases, arthritis, depression, haemophilia, HIV/AIDS, and hypertension. Certain criteria make these diseases particularly suited to DM [see Gillespie 2002”:
– existence of recognised treatment guidelines and agreement as to effectiveness and appropriateness of levels of care;
– known and well-documented problems in therapy that can be addressed by a DM programme;
– large variations in clinical practice and/or drug treatment that can be addressed by a DM programme;
– large numbers of patients with the disease, who are readily identifiable;
– existence of acute events, such as visits to A&E, that are associated with the disease, and which can be prevented by appropriate management;
– existence of outcomes that can be defined and measured objectively; and
– availability of patient education materials and feedback mechanisms to achieve and track changes in behaviour.
In a typical DM programme, plans of care are established, and are geared toward the prevention of secondary symptoms and complications. For example, in diabetes management, the care plan would be designed to manage the disease process itself, as well as known complications such as cardiovascular illness, renal disease, eye disease, and others.

What value is disease management?
In the US system, a major criterion for DM programmes is that they have the potential to demonstrate financial savings for the ultimate healthcare payers, whether that be insurers, employers, or the state. Savings are based on a simple principle: if the savings from decreased use of medical care are greater than the costs of running the programme, the programme has generated a positive return on investment (ROI).
Those US employers and insurers who expected to see savings in the short term from DM have generally been disappointed. For some of these diseases, the costs of care might increase in the short term (as patients become more compliant with testing and drug treatment plans). The savings may emerge much later. For example, early compliance with best-practice protocols for diabetes may have savings many years later by avoiding heart disease and amputations, but seldom leads to immediate savings. It is an investment in reducing future healthcare costs.
There are three main methodologies for calculating the potential cost savings to be gained from DM programmes:
– Comparison of requested services to approved services
During the course of managing a member’s disease, programmes often approve or deny payment for services based upon protocols for managing a disease. Savings are calculated by comparing requested services with approved services.
– Comparison of medical treatment costs for a control group to an intervention group for the same period in time
The medical treatment costs of people enrolled in a DM programme are compared with the medical treatment costs of a group of people with the same chronic disease, but who have not enrolled in the programme. [See Fetterolf et al 2003.”
– Comparison of pre-enrolment medical treatment costs (baseline year) to post-enrolment costs (intervention year)
Total healthcare costs for all patients enrolled for the year prior to enrolment in a DM programme are compared with the same patients’ healthcare costs during subsequent periods after enrolment.
The first method does not translate easily to the UK health environment. The second method is theoretically desirable but hard to achieve in practice. The third method is the most common in the US and holds the most potential for translation to the UK. However, it still presents significant statistical and analytical issues that can confound a ‘true’ estimate of the ROI of a DM programme.

Issues in estimating ROI
There are always issues associated with statistical analyses, ranging from choice of appropriate methods and assumptions to dealing with less-than-perfect data. There are several issues at the core of the analysis of DM outcomes.

Regression to the mean
Members are often recruited into DM programmes when they experience an acute phase of their underlying chronic disease, sometimes requiring hospitalisation. Once the acute phase has passed, they would tend to return to lower levels of use of medical services due to the natural course of the underlying disease process. This trend occurs with or without active intervention by a DM programme.
Figure 1 illustrates the impact of regression to the mean for a US diabetes population under age 65. At the time of a significant event related to the disease (ie hospitalisation or A&E attendance) the mean cost is substantially higher than prior to the acute event. While the post-event mean cost is higher than the pre-event mean, it is markedly less than the mean at the time of that significant event. Results are similar for other typical DM illnesses such as coronary artery disease (CAD) and congestive heart failure (CHF).

Selection bias
There can be significant differences in the use of medical services by the participants in a DM programme compared with those who do not want to enrol. This selection bias may be caused by such things as:
– population-based factors such as geographic, economic, and cultural differences that may affect the prevalence of a condition, as well as the degree of compliance with treatment protocols;
– personal factors such as age, illness severity, and motivation these may influence a person’s willingness to enrol in a DM programme.
In addition, selection bias may be introduced excluding certain groups of people from the DM programme, eg those with substantial co-morbidities, the terminally ill, or those undergoing treatment for cancer or transplants.

Statistical credibility
Disease management populations are a small percentage of the total population, especially in pilot programmes. They are also a high-cost, high-variance population. If certain groups of people with a particular disease are excluded from the DM programme or from the evaluation of the programme, the group will be even smaller and results less credible. Similarly, attempting to adjust for population differences by subdividing the populations (eg by age) can result in group sizes that limit the credibility of the results.

Not all people with the same condition will use the same amount of healthcare services. The variation is pronounced in populations targeted for DM programmes. For example, a DM population with static or declining growth may experience increased average severity owing to the progressive nature of the disease over time and the increased healthcare costs associated with additional severity. As a result, comparing the DM group to a control group without adjusting for the severity change may result in misleading conclusions.
These issues are specific to the evaluation of DM programmes. However, there are other, more familiar actuarial issues that must also be considered in this context. The use of healthcare services will vary between two time periods, independent of the introduction of a DM programme. Among the factors that need to be considered in the evaluation of the ‘true costs’ of a DM programme are:
– Trend If pre-enrolment treatment costs are used as a basis for comparison to post-enrolment treatment costs, they will have to be trended to the current time period. This trend involves more than just the general increase in the use of services and costs over time. Many factors affect trend rates: new technology, new information about existing technology, and changes in medical practice, among others. In any particular year, any one of these factors may have a greater impact on the disease in question than on another disease or on overall healthcare costs, particularly because a specific DM population may access different types of care and use different providers than the average population.
– Changes in health care delivery infrastructure When comparing the costs of a pre-enrolment population to their costs after intervention, it is difficult to isolate the impact of the DM programme from the impact of changes in healthcare delivery. Doctors and hospitals are continually changing their practice patterns and these can drastically change the use of medical services for people with a chronic illness. The more time between the baseline experience and the intervention experience, the greater the impact of changes in care delivery.
Other changes within the delivery system, such as provider payment contracts, hospital contracting cycles, and coding practice shifts, among others, may also have an impact on the services delivered and their costs.

Lessons for the UK
In the UK, the NHS pilot programmes on DM have tended to quote statistics such as: ‘inpatient bed days have been reduced by 40%.’
However, a rigorous analysis of the actual cost of the programme and the cost of the savings has been lacking. In addition, the pilot programmes often involve small numbers of people, and the resultant statistics quoted based on a commensurately small sample.
The US experience shows that a more statistically credible approach can be developed, but the results from these analyses often show that DM programmes increase rather than reduce costs in the short term. NHS or private medical insurance managers who think DM is a panacea to spiralling medical costs should beware. There is no doubt DM programmes have produced better clinical outcomes for the chronically ill. But the jury is still out on the true return on investment.
Aside from the pure cost savings aspect, there have been other measures of success set forth by the DM industry. One of these is ‘quality-adjusted life year’ (QALY). The theory behind QALYs is that there will be an improvement in the quality of life of a patient who is successful in a DM programme and that this can be measured. Using QALYs, the quality and quantity of life following healthcare interventions is estimated and compared to the costs of a DM programme. QALY considers the long-term societal impact of investment in DM. However, because of any potential extra costs on an already over-extended healthcare budget and high employee turnover, most US insurers and employers are reluctant to use QALY to justify investment in a DM programme. The main problem with QALYs is that they do not tell you the potential overall budgetary impact of a DM programme.