Current issues of ACP Journal Club are published in Annals of Internal Medicine


Review: Prompted follow-up by specialists and GPs produces similar outcomes for diabetes mellitus

ACP J Club. 1998 May-June;128:57. doi:10.7326/ACPJC-1998-128-3-057

Source Citation

Griffin S, Kinmonth AL. Diabetes care: the effectiveness of systems for routine surveillance for people with diabetes. In: The Cochrane Database of Systematic Reviews. The Cochrane Library. Oxford: Update Software; 1998, Issue 1.



To compare the effectiveness and cost, using meta-analysis, of primary care from general practitioners (GPs) with secondary care from hospital-based specialists for review and surveillance of complications of diabetes mellitus.

Data sources

Studies were identified using MEDLINE, CINAHL, National Research Register, Psychlit, Healthstar, EMBASE, CRIB, and Dissertation Abstracts (from year of inception to end of 1996) with terms relating to general practice, primary health care, and diabetes. The Cochrane Diabetes Group Trials Register, the Cochrane Library, and bibliographies of relevant studies were also searched.

Study selection

Studies were selected if diabetic patients were randomly allocated to either follow-up by GPs (some of whom provided shared care with a specialist) or hospital-based specialists.

Data extraction

Data were extracted on setting, duration of follow-up, randomization method, patient characteristics, intervention, and outcomes (mortality, metabolic control, cardiovascular risk factors, quality of life, functional status, satisfaction with care, hospitalizations, cost and completeness of screening, and diabetes complications).

Main results

5 studies (1058 patients) met the inclusion criteria. 4 studies reported patient age and sex (overall mean age 58 y, 56% men). Analysis was done by intention to treat. Only 1 trial had a duration > 2 years. In 2 studies, routine care from a GP was compared with care from a specialized diabetes clinic; in 2 other studies, prompted care from a GP was provided by a central computer system; and in 1 study, both approaches were used. Overall, mortality (odds ratio [OR] 1.75, 95% CI 1.12 to 2.75) and losses to follow-up (OR 3.21, CI 2.27 to 4.56) were higher among patients who received care from a GP. Patients who received prompted care from a GP were reviewed more often, had more frequent testing of glycosylated hemoglobin levels, and were more likely to be referred to a podiatrist (OR 2.51, CI 1.54 to 3.97). However, these patients were less likely to be referred to a dietitian (OR 0.61, CI 0.40 to 0.92) than those who received hospital-based clinic supervision. No difference existed between patients receiving prompted GP and hospital-based specialist care for metabolic control, blood pressure, and mortality. Data on health service costs varied among studies. Most studies did not examine quality of life, cardiovascular risk factors, functional status, or diabetes complications.


Routine care from a general practitioner is associated with higher mortality and greater losses to follow-up than hospital-based specialist care. Prompted care from a general practitioner is associated with more frequent reviews and testing of glycosylated hemoglobin levels, more podiatrist referrals, and fewer dietitian referrals.

Sources of funding: Primary Medicine Care Group, University of Southampton and The Wellcome Trust UK.

For article reprint: Not available.


The care of patients with diabetes is increasingly becoming the jurisdiction of the GP and the primary health care team (PHCT). We know that complications can be reduced by regular surveillance and screening; thus, evidence of effective care by the PHCT is important for us to have confidence in and be able to encourage this process.

The Cochrane review by Griffin and Kinmonth of 5 trials done in Australia and Britain used the strict inclusion criterion of random allocation of patients with type 1 or 2 diabetes to either follow-up from hospital-based specialists or GPs and the PHCT. It includes 2 trials from the 1980s that had no recall systems and 3 from the 1990s that had careful prompting or recall. The prompted group shows that the PHCT is able to care for patients with diabetes if certain criteria are met. Those patients were seen more often, had glycosylated hemoglobin levels checked more often, and were more likely to be referred to a podiatrist but not to a dietitian.

The lessons to be learned from this review are that recall systems for all patients with diabetes mellitus must be mandatory, whether used centrally or at the practice level, and that all providers of routine surveillance must agree on well-defined protocols for their diabetic patients.

Education, not only initially but also on a continuing basis, is necessary for all health care professionals who provide diabetic care—this includes consultants and junior hospital medical staff, GPs, practice and community nurses, dietitians, and podiatrists, preferably working together (1). The provision of care in a primary care setting also requires adequate financial, clinical, and administrative support.

The number of patients who have diabetes is increasing. The reduction of complications is important to both patients and health services providers. Primary care must deliver high-quality diabetic care to satisfy patient needs.

Tim Carney, MD
Burn Brae Medical GroupHexham, Northumberland, UK


1. Carney T, Helliwell C. Effect of structured postgraduate medical education on the care of patients with diabetes. Br J Gen Pract. 1995;45:149-51.