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

Implementing better diabetes care

Problems and changes

There is quite a lot of evidence about clinical interventions to improve diabetes care, as well as evidence about management interventions that can improve care and patient satisfaction. We know that improved care leads to better glycaemic control, and that, together with appropriate control of weight, cholesterol, and blood pressure, results in improved outcomes.

More problematic is the business of putting it all together in an overall package of care to deliver the goods. Here we move from randomised to observational studies, and from the hundreds and thousands to the tens of thousands or millions of patients. A report from Israel [1] demonstrates that putting it all together begins to deliver the goods.

Problems and changes

The setting was a healthcare services organisation serving a quarter of the population of Israel (about 1.5 million people), with almost 3% diagnosed as diabetic. Issues with structure of care provision, the problems caused, and the solutions used in a new care programme are shown in Table 1.

Table 1: Problems and solutions around implementing better diabetes care

Structure issues
Self-referral to diabetic clinic Overloaded clinics
Imbalance between need and delivery of specialist treatment
GP referral only
Clinic outreach for specialist treatment by need, via GP
Absence of guidelines No agreed standard of care Multidisciplinary consensus programme development and dissemination
Absence of central monitoring No information for decision-making
No feedback
Automated registry provided feedback at all levels
Split responsibility for care Little communication between providers
Difficulties in coordinating care

Specialists unwilling to discharge patients to GP
Centralised monitoring
Diabetic clinic responsible for patient care, coordination, and education, and for patient empowerment
More patients with complex problems treated by specialists

There were two main thrusts to the programme:

Each clinical diabetologist was instructed to spend 20% of their working time on different aspects of disease management instead of direct patient care, time being created by discharging patients not on insulin to primary care. Education, feedback, patient support groups and the working of the registry were the core features of the programme.

The population was the 22,000 members continuously in the diabetes register from 1999 to 2004. The intervention began in January 2000.


Urine collections for micro-albumin analysis, eye tests, and prescription of ACE-inhibitors increased with the new programme (Figure 1), as did testing of glycosylated haemoglobin and LDL-cholesterol from 74% in 1999 to 88% in 2004. Other process indicators, like blood pressure and BMI were not available.

Figure 1: Process indicators in implementing better diabetes care

Mean levels of glycosylated haemoglobin fell from 8.1% in 1999 to 7.8% in 2004. An absolute reduction of 2% or more occurred in 6% of patients, and of 1% or more in 16%.

Mean levels of LDL-cholesterol fell from 3.3 mmol/L in 1999 to 2.9 mmol/L in 2004. In 12,000 patients having LDL-cholesterol tested each year, the average absolute reduction was by 0.34 mmol/L.

The number of patients seen at diabetic clinics rose by 61% with only a 23% increase in work hours. Resources increased less than the increase in diabetic numbers.


The study followed a large cohort over five years, before and after introduction of a reformed programme of diabetic care. Measured indicators of the care process demonstrated large improvements, and measured indicators of benefit improved. Mean glycosylated haemoglobin levels dropped by 4% over five years, in contrast to the increase of 7% usually seen as patients age by five years.

Keys to success were good information technology, communal ownership of the programme, and being able to step back and make a big change. The diabetes register and common ownership maximised care and spotted when patients needed specialist care. Evidence, plus thoughtful management, provided better care, at a lower cost per patient. Impressive stuff.


  1. AD Heymann et al. The implementation of managed care for diabetes using medical informatics in a large preferred provider organisation. Diabetes Research and Clinical Practice 2005 (e-publication ahead of print).

previous or next story