Skip navigation
Link to Back issues listing | Back Issue Listing with content Index | Subject Index

Improving care for chronic illness


Chronic illness is common. Depending where you go to access statistics, it is as many as a third to a half of people in the USA, 90 to 125 million people with at least one chronic illness, with a fifth (50 million in the USA) with two or more chronic illnesses. Chronic illness is very expensive, and asthma, diabetes, heart disease, hypertension, and mood disorders account for up to half of all healthcare expenditures in more developed countries. Use of healthcare resources by someone with a chronic illness can be five times higher than by a healthy person.

Much chronic illness is in the community, and delivering the right care to the right patient in the right way could have significant benefits. In the USA a Chronic Care Model (CCM) has been developed as a framework for improving chronic disease care in the community. The CCM has six elements for providing high quality services (Table 1), four of which are directly related to care delivery. The question is whether these elements do contribute to better care, and a meta-analysis [1] suggests that they do.

Table 1: Elements of the Chronic Care Model

General Specific
Delivery system design Care management roles
Team practice
Care delivery coordination
Proactive follow-up
Planned visit
Visit system change
Self-management support Patient education
Patient psychosocial support
Self-management assessment
Self-management resources
Collaborative decision-making
Guidelines available to patients
Decision support Insitutional guidelines or prompts
Provider education
Expert consultation support
Clinical information systems Patient registry system
User information for care management
Feedback of performance data
Community resources For patients
For community
Health care organisation Leadership support
Provider participation
Coherent system improvement and spread


Four clinical areas were of interest: asthma, congestive heart failure, depression, and diabetes. Systematic reviews and meta-analyses of chronic illnesses and electronic searches were used to find relevant studies. Included studies could be randomised or observational, as long as the intervention used at least one CCM element. Outcomes chosen were continuous or dichotomous clinical variables, some prespecified, like emergency department visits for asthma or congestive heart failure, or HbA 1c for diabetes. Quality of life information was also sought, and process of care indicators, like number of patients receiving appropriate medicines, or being tested for HbA 1c. Pooled estimates for interventions versus controls were then calculated in each of these areas, for each of the clinical conditions and overall.


They found 112 studies, 93% of which were randomised. Most were published since 1999, so were recent. Most (107) were in an outpatient setting looking at one or two interventions, and half had follow ups of six months or longer. Of the randomised trials, few had good quality scores, but were adequate given that blinding was next to impossible. Only a third had reasonably large numbers (>200) patients.

Results are summarised in Table 2. For most of the outcomes over most of the conditions, use of a care package with at least one of the CCM interventions delivered a significantly better result. For instance, for diabetes this was equivalent to a 0.3% to 0.5% reduction in HbA 1C. Results were even better when restricted to randomised trials with better quality scores. Most information was available for interventions that included elements of delivery system design, decision support, or self-management support, and subgroup analyses for these were generally significant.

Table 2: Analysis by element and overall, shaded areas showing statistical improvement

Outcome variable
Quality of life
Process of care
Congestive heart failure


This all sounds a bit management-speak, and, to some extent, it is. A word of caution also, because none of the supporting links from the paper worked on the day that Bandolier tried. But this should not be dismissed. Material on the Internet shows that some primary care settings in the USA have had terrific results, improving patient care while reducing costs – what every decent care pathway should deliver. And delivery of healthcare is a complex business, not much studied. Perhaps we suffer from too much policy-based evidence, rather than evidence-based policy. Certainly this is worth a detailed look for those working in chronic care delivery.


  1. AC Tsai et al. A meta-analysis of interventions to improve care for chronic illness. American Journal of Managed Care 2005 11: 478-488.

previous or next story