Article Text


Diabetes care provision and glycaemic control in Northern Ireland: a UK regional audit
  1. C R Cardwell1,
  2. C C Patterson1,
  3. M Allen2,
  4. D J Carson2,
  5. on behalf of the Northern Ireland Paediatric Diabetes Study Group
  1. 1Department of Epidemiology & Public Health, The Queen’s University of Belfast, UK
  2. 2Department of Child Health, The Queen’s University of Belfast, UK
  1. Correspondence to:
    Mr C R Cardwell
    Department of Epidemiology & Public Health, The Queen’s University of Belfast, Grosvenor Road, Belfast BT12 6BJ, Northern Ireland, UK;


Aims: To assess the care received, compared to national guidelines, and to investigate factors associated with glycaemic control in children and adolescents with type 1 diabetes attending clinics in Northern Ireland.

Methods: An audit of the care provided to all patients attending 11 paediatric diabetes clinics commenced in 2002. A research nurse interviewed 914 patients completing a questionnaire recording characteristics, social circumstances, and aspects of diabetes management, including the monitoring of complications and access to members of the diabetes team. Glycaemic control was measured by glycosylated haemoglobin (HbA1c), determined at a DCCT aligned central laboratory.

Results: The average HbA1c concentration was 8.8% (SD 1.5%), with 20% of patients achieving recommended HbA1c levels of less than 7.5%. In the year prior to the audit, 76% of patients were reviewed by a diabetes specialist nurse and 42% were tested for microalbuminuria. After adjustment for confounding factors, better glycaemic control was identified, particularly in patients who had attended exactly four diabetes clinics in the previous year, were members of the patient association Diabetes UK, and lived with both natural parents.

Conclusions: In Northern Ireland only a minority of patients achieved recommended HbA1c levels. Furthermore, children and adolescents with diabetes were reviewed by fewer specialists and were less intensively monitored for microvascular complications than recommended. There was evidence of better control in children who were members of Diabetes UK, suggesting that parental attitude and involvement could lead to benefits.

  • type 1 diabetes mellitus
  • glycaemic control
  • adolescence
  • clinical standards

Statistics from

Good glycaemic control in children and adolescents with type 1 diabetes has been associated with better quality of life1 and reduced or delayed development of long term complications.2,3 The structure of an appropriate multi-professional team was recommended in the UK by The principles of good practice for the care of young people with diabetes.4 The provision of these services has been surveyed on three occasions between 1988 and 1999,5–7 and although few clinics achieve all standards, results have improved over time. However, information was obtained from questionnaires completed by paediatricians without reference to individual patients or their medical records.

The provision of a motivated multi-disciplinary team of diabetes specialists8,9 is thought to be vital in attaining optimal glycaemic control. However, international10 and national cross-sectional studies11–14 continue to report glycosylated haemoglobin (HbA1c) results for the majority of patients above the levels desirable to minimise the risk of complications in adulthood. Various factors have been identified which are associated with glycaemic control,10,12,13,15–20 but further work is required.21

The goals of this audit were to identify all patients attending paediatric diabetes clinics in Northern Ireland and obtain information about the care received over the previous 12 months directly from patients and/or their parents and from medical records. The audit also determined HbA1c levels for these patients and investigated factors associated with glycaemic control.


Between April 2002 and June 2003 all children and adolescents attending 11 paediatric diabetes clinics in Northern Ireland were invited to participate. An estimated ascertainment rate was calculated as the percentage of patients on the Northern Ireland childhood type 1 diabetes register22 on 1 January 2002, under 15 years of age, still resident in Northern Ireland, and not attending adult clinics, who were included in the audit. A single research nurse collected information during routine outpatient visits. Families missing their appointments were interviewed at a subsequent appointment. For each patient a structured questionnaire was completed recording characteristics (including age, weight, height, duration of diabetes, postcode of usual address, clinical history, and family history of autoimmune disease), aspects of diabetes management (including insulin dosage, blood glucose monitoring, attendance at clinics, occurrence of severe hypoglycaemia, and other complications), and social factors (including social class and family circumstances). This questionnaire also recorded contacts with healthcare specialists and the monitoring of complications over the previous 12 months; simultaneously, this information was independently retrieved from hospital case notes. Hypoglycaemic episodes were considered severe if intramuscular glucagon or intravenous glucose was administered. Finally, a 50 μl blood sample was taken and HbA1c measured at the Special Investigations Haematology Laboratory, Royal Group Hospitals Trust, Belfast by a DCCT aligned HPLC method.

Audit criteria were adapted from the third national UK survey7 and from the International Society for Pediatric and Adolescent Diabetes recommendations.23 Analysis of audit criteria were conducted in patients whose diabetes had been diagnosed for at least 12 months prior to interview, as many criteria refer to care in that period.

Analysis of glycaemic control, severe hypoglycaemic episodes, and readmissions for diabetic ketoacidosis (DKA) were restricted to patients interviewed prior to their 16th birthday as coverage of older patients was considered less complete. Patients whose diabetes had been diagnosed for less than 12 months who may be in a phase of partial remission were also excluded. A deprivation score was calculated for each postcode based on the Carstairs index, derived from four 1991 census indicators: unemployment, car ownership, social class, and household crowding.24

Statistical analysis

Height, weight, and BMI were converted into standard deviation scores (SDS) to take account of children’s gender and age using the 1990 British Growth Standard25 and then divided by quintiles into fifths. The mean and standard deviation of HbA1c was obtained for each category of potential explanatory variables and compared using t tests and one way analysis of variance. Multiple regression analysis was conducted controlling first for sex, age, and duration (in categories), as these were factors over which clinics had no control, and then for other factors on a variable-by-variable basis. Rates of DKA and severe hypoglycaemic episodes were calculated and a χ2 test used to compare severe hypoglycaemia by HbA1c level (categorised as >7.5% or ⩽7.5%).

Statistical analyses were performed using STATA release 8.0 (Stata Corporation, College Station, Texas).


There were 914 patients included in the audit, all Caucasian, and none with type 2 diabetes. Other patient characteristics are shown in table 1. The estimated ascertainment rate was 97.4%.

Table 1

 Characteristics of all patients included in the study attending paediatric diabetes clinics in Northern Ireland (n = 914)

There were 778 patients whose diabetes had been diagnosed for at least one year. Review and monitoring of complications by healthcare specialists are shown in table 2.

Table 2

 Compliance with audit criteria, taken from hospital charts, of patients attending paediatric diabetes clinics in Northern Ireland whose diabetes had been diagnosed for at least one year (n = 778) and for children over 11 years of age (n = 452)

Mean HbA1c was 8.8%, standard deviation 1.53%, in the 95% (621/651) of patients, interviewed prior to their 16th birthday whose diabetes had been diagnosed for at least one year, for whom this was available. Comparison of HbA1c levels in subgroups of these 621 patients are shown in table 3 (sociodemographic factors) and table 4 (clinic related factors). These tables show comparisons with a reference category for each patient characteristic after controlling for sex, age, and duration.

Table 3

 Comparison of HbA1c results by patient characteristics in patients interviewed prior to their 16th birthday whose diabetes had been diagnosed for at least one year (n = 621)

Table 4

 Comparison of HbA1c results by aspects of diabetes management in patients interviewed by the research nurse prior to their 16th birthday with duration of diabetes over 1 year (n = 621)

In table 3 the unadjusted analysis shows that age, pubertal status, duration of diabetes, SDS height, and SDS weight were all significantly associated with HbA1c level. The association with age was independent of sex and duration (p < 0.0001) and the association with duration was independent of age and sex (p = 0.01). Patients who were not living with both their natural parents had a significantly higher HbA1c by an estimated 0.55% (95% CI 0.22% to 0.88%) after adjusting for sex, age, and duration. Although parental occupation was significantly associated both before (p = 0.0004) and after adjustment (p = 0.005), the effect was attributable to high HbA1c levels in those patients where the head of the household’s occupation could not be classified (for example, never worked, long term unemployed, or armed forces).

In table 4 the crude analysis shows that the number of blood glucose tests per day, number of HbA1c tests, number of doses of insulin, and total insulin dose were all associated with HbA1c level. However, some of these associations could be attributed to other factors. After adjustment for sex, age, and duration the associations between the number of doses of insulin per day and the total insulin dose were no longer significant. After additionally adjusting for the attended diabetes clinic the associations between both number of blood glucose tests per day (p = 0.21) and number of HbA1c tests (p = 0.10) were no longer significant.

The number of diabetes clinic attendances was associated with HbA1c level both before and after adjustment. Specifically, after adjusting for sex, age, and duration, those patients attending less than four clinics in the last year had a significantly higher HbA1c level by an estimated 0.42% (95% CI 0.14% to 0.70%) compared to those attending exactly four clinics, while those attending more than four clinics had a higher HbA1c level by an estimated 0.32% (95% CI 0.03% to 0.61%) compared to those attending exactly four clinics. Members of Diabetes UK had a significantly lower HbA1c by an estimated 0.49% (95% CI 0.26% to 0.73%), after adjustment for sex, age, and duration. Other factors investigated for which no association between HbA1c was detected included season of visit and time from home to clinic. The model containing sex, age, duration, clinic attendance, natural parents, and Diabetes UK membership explained 13% (R2 = 0.13) of patient-to-patient variation in HbA1c levels; adding clinic to this model raised the figure to 21%, indicating that clinic-to-clinic differences in HbA1c values remained that could not be explained by the available patient characteristics.

There were 92 patients (14%), interviewed prior to their 16th birthday whose diabetes had been diagnosed for at least one year, who reported 191 severe hypoglycaemic episodes in the previous year, which corresponded to 29.5 episodes per 100 person-years. In contrast, the hospital chart recorded a total of 13 severe hypoglycaemic episodes, a rate of 2.0 per 100 person-years, occurring in 1% (8/648) of patients. There was no significant difference (p = 0.27) in the prevalence of severe hypoglycaemia by HbA1c; patients with HbA1c ⩽7.5% experienced a rate of 25.0 episodes per 100 person-years compared with 31.1 in patients with HbA1c >7.5%. There was little difference between the chart and interview record of readmissions for DKA; overall the chart recorded 7.1 DKA episodes per 100 person-years, which occurred in 6% (36/649) of the children.


This is the first audit of the care provided to children and adolescents with type 1 diabetes in Northern Ireland. Glycaemic control, mean HbA1c 8.8%, was roughly comparable to other similar national audits from Scotland,13 France,12 and two audits from Denmark11,14 (mean HbA1c 8.9%, 9.0%, 9.1%, and 8.7%, respectively). However, the small percentage, 20%, of patients achieving recommended HbA1c levels,23 ⩽7.5%, is a cause for concern.

The lack of patients receiving an annual review by all members of the paediatric diabetes specialist team may reflect the failure to meet healthcare staffing recommendations. In particular, throughout Northern Ireland there were only 2.5 whole time equivalent (WTE) diabetes specialist nurses in post compared with the 9 diabetes specialist nurses recommended26 for a clinic population of 914. Similarly, the low proportion of patients reviewed by a clinical psychologist may also reflect staff shortages rather than a lack of need for psychological input among children with persistently poor glycaemic control.27

Monitoring of children and adolescents for microvascular complications, particularly microalbuminuria and retinal examinations, fell below recommendations.23 This study shows how the deficiencies in diabetes services, based on a recent UK survey of consultant paediatricians,7 translate to deficiencies in the care received by patients. Importantly, this is the first study that is likely to accurately reflect the levels of care provided in the UK as information was directly recorded from parents/carers and was confirmed by chart review. In contrast, the previously mentioned survey relied on the perceptions of consultant paediatricians who were responsible for providing diabetes services. Two other UK audits,13,28,29 to our knowledge the only UK audits of similar size and ascertainment rate, did not present information on care received directly recorded from the patient/carer.

The observation that less than 10% of children stating they had experienced severe hypoglycaemic episodes in the past year had this recorded in their hospital chart suggests that these episodes may not be receiving adequate attention at clinic visits. Furthermore, this result indicates that comparisons of severe hypoglycaemia between centres must be interpreted with caution.

The observed associations between HbA1c level and age and diabetes duration and the lack of association with gender, were mostly consistent with previous large cross-sectional studies.10,12,13 Earlier studies have also observed HbA1c associations with total insulin dosage12,13,16,18,20 and the number of insulin doses per day,12,13 but in this study these associations could be explained by the duration of diabetes and the age of the patients. However, poorer control was noted in children prescribed four doses of insulin per day, which may reflect that this insulin regime is used for children with poor control or that it may be associated with poor compliance. The evidence supporting the concept that increased blood glucose testing improved control, a consistent finding in previous studies,12,16,18 was less conclusive. An increase in the number of blood glucose tests was associated with a decrease in HbA1c, but after adjusting for the attended diabetes clinic this effect was no longer apparent, suggesting that it may be explained by other clinic practices.

Glycaemic control was independently associated with Diabetes UK membership, diabetic clinic attendance, and living with both natural parents. The reduction in HbA1c with Diabetes UK membership suggests that parental attitude and involvement in the disease could lead to benefits in control. This study confirms an association between infrequent diabetes clinic attendance and worse control that has been noted previously.16,17,19 However, caution should be exercised when interpreting the results in the most frequent category of attendance, as increased attendances could be a consequence rather than a cause of worse control. The worse control of children whose parents have separated has been observed in various studies;13,18,19 the consistency of this finding indicates that single parents may merit assistance to meet the challenges involved in attaining good control.

In addition to previously mentioned advantages, this study also benefited from a relatively large sample size, a high ascertainment rate, and a detailed questionnaire on established and potential risk factors of poor control. This allowed not only the testing of previous hypotheses but also the adjustment for other risk factors of poor control. This study does have limitations; the cross-sectional design allows the association of factors with good control but not the ability to determine causation which would require a prospective study. Furthermore, the 21% of HbA1c variation that was explained by available variables was similar to a Scottish study,13 which accounted for 16% of HbA1c variation, but indicated that knowledge of factors which establish good control remains incomplete.

In conclusion, the percentage of children and adolescents achieving recommended levels of glycaemic control was disappointing. Furthermore, annual reviews by healthcare specialists and monitoring of microvascular complications fell below recommended levels. Some of these failures may reflect staff shortages, in particular a lack of diabetes specialist nurses and clinical psychologists. There was evidence of better control in children who were members of Diabetes UK, who lived with both natural parents, and who attended diabetes clinics regularly.

Embedded Image


The audit was undertaken with the support and participation of the following members of the Northern Ireland Paediatric Diabetes Study Group: M Allen, T Blair, R Blakely, M Brock, C Cardwell, D Carson, C Corkey, P Cosgrove, N Craig, M Doherty, M Allen, C Gaston, S Griffin, V Gleadhill, C Halahakoon, L Irwin, O McGlone, J Newell, C Patterson, M Quinn, M Rollins, C Stewart, and H Tennet.


View Abstract


  • Funding for the study was provided by the Regional Multi-professional Audit Group, Department of Health, Social Services, and Public Safety. CRC was in receipt of a Department of Education and Learning grant.

  • Competing interests: none declared

  • Ethics approval: ethical approval was obtained from Queen’s University Belfast Research Ethics Committee

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.