Article Text

Download PDFPDF

Delayed diagnosis in type 1 diabetes mellitus
  1. P C B Sundaram,
  2. E Day,
  3. J M W Kirk
  1. Birmingham Children’s Hospital, Birmingham, UK
  1. Dr Jeremy Kirk, Paediatric Endocrinology and Diabetes, Birmingham Children’s Hospital, Steelhouse Lane, Birmingham B4 6NH, UK; jeremy.kirk{at}


Children with suspected type 1 diabetes mellitus (T1DM) should have same day referral to a paediatric diabetes team. 99 children (54 male; median age 10.5 years, range 0.9–15.9 years) were diagnosed with T1DM at our hospital between January 2004 and June 2007. 27 (27.2%) presented in diabetic ketoacidosis (DKA). 37 (37.3%) required hospital admission, while the rest had ambulatory management. In 21 (21.2%) children, diagnosis was delayed >24 h (median 3.0 days, range 1–14 days) due to missed diagnosis at the local hospital (four) or by the general practitioner (seven), arranging a fasting blood glucose test (nine) and outpatient appointment requested via fax (one). Children with delayed diagnosis presented more frequently in DKA (52.3% vs 20.5%, p<0.01), with a higher median presenting HbA1c (12.3% vs 10.9%, p<0.05). There were no differences in age and sex between the delayed diagnosis and immediate referral groups. Healthcare providers need to be aware of the importance of immediate referral of children newly diagnosed with T1DM.

View Full Text

Statistics from

The National Institute for Health and Clinical Excellence (NICE) recommended in 2004 that children with suspected type 1 diabetes mellitus (T1DM) should be offered immediate (ie, same day) referral to a paediatric diabetic multi-disciplinary team that has the competencies needed to confirm the diagnosis and to provide immediate care.1 Failure to consider the possibility of a diagnosis of T1DM in spite of classical symptoms and atypical presentation may lead to delayed diagnosis.

The aim of our study was to assess any delay in the diagnosis and management of children newly diagnosed with T1DM after seeing a health care professional, and whether this delay had any implications on presenting features and outcome.


Children diagnosed with T1DM between January 2004 and June 2007 at Birmingham Children’s Hospital were identified from our paediatric diabetic database (Twinkle). Details of age, sex, ethnicity, and mode of presentation, duration of symptoms and blood glucose, pH and HbA1c levels at presentation were obtained from the patient case notes. Both the referral letter from the general practitioner (GP) and the hospital case notes were analysed to see if the child had been seen by a medical professional 24 h before the diagnosis. A gap of more than 24 h between the initial presentation to a primary or secondary care provider and referral to the multi-disciplinary diabetic team was considered to be delayed diagnosis. Comparison was made between children with delayed diagnosis and those with immediate referral. Diabetic ketoacidosis (DKA) was defined as a venous pH <7.3 or bicarbonate <15 mmol/l in the presence of hyperglycaemia (blood glucose >11 mmol/l) and ketonaemia and/or ketonuria.

Graph Pad Prism and SPSS were used for analysis of the results. Fisher’s exact test was used for comparing the categorical variable between the delayed diagnosis and immediate referral groups, while the Mann–Whitney test was used for continuous variables.


Ninety nine children (54 male (55%), 45 female) were diagnosed with T1DM between January 2004 and June 2007. The median age at diagnosis was 10.5 years (range 0.9–15.9 years). Seventy (70.7%) children were from a UK white background and 29 (29.3%) were from other groups including South Asian (17), African (6), East European (2) and mixed race (4).

Twenty seven (27.2%) children presented with DKA at diagnosis. This was further sub-classified as mild (6), moderate (14) and severe (7) according to whether the initial pH was less than 7.3, 7.2 or 7.1, respectively.2 Thirty seven (37.3%) children required hospital admission; the remainder had ambulatory management from diagnosis under the diabetes home care team.

Twenty one (21.2%) children were documented to have a delay in diagnosis. The median duration of delay was 3 days (range 1–14 days). The commonest reason for delay was because the referring health professional was arranging a fasting blood glucose test after T1DM was suspected (nine children). Other reasons were the diagnosis was missed by the GP (seven) or the hospital doctor (six), and the GP was trying to arrange a hospital appointment via fax (one).

Those children whose diagnosis was missed by the GP had osmotic symptoms of T1DM subsequently documented in the hospital notes, and three also had candidiasis (thrush). Although three children had presented to hospital with abdominal pain and vomiting, urine examination was not performed. The other child presented with chest pain and hyperventilation.

Comparisons between children with delayed diagnosis and immediate referral are shown in table 1. There were no differences in the age and sex of children between the delayed diagnosis and immediate referral groups. However, there were more children from a non-white background in the delayed diagnosis group when compared to the other group (42.8% vs 25.2%), although this was not statistically significant. The most striking finding in our study was that children who had a delayed diagnosis had a significantly increased risk of DKA at presentation (52.3% vs 20.5%, p<0.05). This group also had a statistically significant higher median HbA1c level and lower median pH at diagnosis. They also had a higher blood glucose level, although this did not achieve statistical significance. Analysis of factors involved in delayed diagnosis using a stepwise forward and backward logistic regression model did not show an effect of age, gender, ethnic origin or duration of symptoms.

Table 1 Comparison between children with delayed diagnosis and immediate referral


This study demonstrates that children with a delayed diagnosis of T1DM are more likely to present with DKA, thus confirming the importance of the NICE recommendation of immediate referral of children with suspected T1DM to a multi-disciplinary diabetic team. We also noted that the odds of presenting with DKA in a child with delayed diagnosis were 5.5 times higher compared to a child presenting with DKA with no delay in diagnosis. The commonest reason for the delay was because a fasting blood glucose test was being arranged. In the presence of osmotic symptoms, however, a random blood glucose level of >11.1 mmol/l is sufficient to make the diagnosis.3 Our study also showed that 42.8% of children in the delayed diagnosis group were from ethnic minorities compared to 25.6% in the immediate referral group, although this did not reach statistical significance, probably due to the small numbers. The overall incidence of DKA at diagnosis was 27.2%, which was similar to that found in other studies.2 The incidence of DKA at presentation was higher in ethnic minorities (41.3%) than in children from a white background (21.4%), which concurs with earlier findings from our unit4 showing that 48% of Asian children with newly diagnosed T1DM presented with DKA. Van Laar et al5 have shown that non-Western immigrant children were likely to be sicker at first presentation with T1DM when compared to Dutch children, suggesting a delay in diagnosis. The possible explanations may be difficulties in recognising the symptoms due to language and cultural barriers, and lack of awareness of T1DM in ethnic minorities.

The prevention programme for DKA in children carried out in Parma, Italy showed that information on diabetes displayed on posters targeted at teachers, parents, students and paediatricians reduced the incidence of DKA in newly diagnosed children.6 Our paper also strengthens the need for educating not only medical professionals but also the general public about the symptoms of diabetes in children to prevent delay in diagnosis.


Our study has confirmed the need for immediate referral of children with suspected T1DM, as delays in diagnosis increase the risk of DKA. Primary and secondary care providers need regular and ongoing education regarding diagnosis of T1DM and the need for immediate referral.


We would like to thank Professor Tim Barrett and Dr Nick Shaw for enabling us to use data on their patients, Lesley Porter for loading the Twinkle database, and Dr Sarah Ehtisham and Dr Nicola Crabtree for assistance with the statistical analysis.


View Abstract


  • Funding: None.

  • Competing interests: None.

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.