DIABETICMedicine DOI: 10.1111/dme.12779
Research: Care Delivery Effectiveness of a multidisciplinary team approach to the prevention of readmission for acute glycaemic events D. Simmons1, S. Hartnell1, J. Watts1, C. Ward1, K. Davenport1, E. Gunn2 and A. Jenaway2 1 Institute of Metabolic Science, Cambridge University Hospitals NHS Foundation Trust and 2Cambridgeshire and Peterborough NHS Foundation Trust, Cambridge, UK
Accepted 8 April 2015
Abstract Aims To describe the effect of a combined diabetes specialist/mental health team approach to prevent readmissions for acute glycaemic events among patients with diabetes. Methods Consecutive patients with diabetes, readmitted to a single hospital for an acute glycaemic condition, were offered one or more diabetes (including assessment, education, medication, technology use and intensive support) and mental health (including assessment, training and therapies) interventions. The pilot service took place over 11 months, with the preceding 24 months and subsequent 8 months serving as control periods.
Of the 58 patients admitted, 50 had Type 1 diabetes and were from within the hospital catchment area, and were discharged home. Of these, 32 (64%) had a pre-existing mental health issue and 14 (28%) had a complex social situation. In all, 96% of patients were met as an inpatient by a team member, and 94% accepted at least one intervention. The mean SD number of admissions per patient/month dropped from 0.12 0.10 to 0.05 0.10 (P < 0.001) during the intervention, increasing, once the intervention ended, to 0.16 0.36 (P = 0.002). The mean sd length of stay similarly decreased and increased (0.6 0.9 to 0.2 0.7 days; P < 0.001 to 0.006) to 0.6 1.4 days (P = 0.003) per patient/month) across the three periods, as did the mean SD tariff paid per patient/ month (₤258.0 374.0 vs ₤92.1 245.0 vs ₤287.3 563.8; P < 0.001 and P = 0.018, respectively). The mean sd HbA1c level dropped from 99 22 to 92 24 mmol/mol (11.2 4.2% vs 10.6 4.3%; P = 0.014) but did not increase after the intervention [89 26 mmol/mol (10.4 4.5%)]. Results
Conclusions The cost and long-term risks of hospitalization among patients with Type 1 diabetes and recurrent admissions can be reduced by a combined specialist diabetes/mental health team approach.
Diabet. Med. 32, 1361–1367 (2015)
Introduction Both hypoglycaemia and diabetic ketoacidosis, although preventable, remain life-threatening complications of Type 1 diabetes [1]. Historically, patients with recurrent admissions for hypoglycaemia and diabetic ketoacidosis have been termed ‘brittle’, and experience poorer quality of life, more complications and premature death [2–4]. Since the early descriptions of brittle diabetes, new approaches to ketoacidosis/hypoglycaemia management have been introduced to reduce inpatient morbidity/mortality [5], newer insulin formulations/delivery systems have been developed [6] and better patient education for self management of Type 1 diabetes [e.g. the Dose Adjustment For Normal Eating Correspondence to: David Simmons. E-mail: [email protected]
ª 2015 The Authors. Diabetic Medicine ª 2015 Diabetes UK
(DAFNE) programme] has been introduced [7]. These developments have been associated with less hypoglycaemia and ketoacidosis [7,8]. To be effective, however, such developments have to be accessible to patients. A recent study showed that, among African-American people in Atlanta, USA [9], 55% of diabetic ketoacidosis cases were readmissions, with insulin discontinuation the leading precipitating cause, particularly as a result of financial issues. Few studies report the characteristics of those who experience recurrent admissions in settings where these modern approaches to diabetes management are available, and where access to insulin should not be a problem, e.g. the National Health Service (NHS) in the UK. Similarly, there are few reports of intervention studies among adults with recurrent admissions for acute glycaemic events [10], although education, case management and
1361
DIABETICMedicine
What’s new? • Patients with recurrent hospitalization for acute glycaemic events mainly have Type 1 diabetes and usually also have mental health and/or social issues. • Multidisciplinary approaches that combine diabetes, mental health, case management and new technology approaches to assist and support patients with recurrent hospitalization can substantially reduce hospital readmissions, length of stay, cost of admissions and HbA1c levels. • Although some patients with recurrent admissions can achieve independence after intervention, the majority require ongoing support. • Patients with alcohol and/or drug-dependency are particularly challenging. psychological interventions have been recommended to prevent acute events among children and adolescents with recurrent admissions [11,12]. A local audit of 473 patients admitted over 12 months (2010/2011) with ketoacidosis, hyperglycaemia or hypoglycaemia included 168 readmissions (14/month). In the present study, we describe the effectiveness of an 11-month diabetes case management service for adult patients with recurrent hospital attendance, which was designed to reduce hospital bed days attributable to glycaemia-related admissions.
Methods Patients
The study included consecutive patients admitted to Addenbrookes Hospital, Cambridge, with either hyperglycaemia or hypoglycaemia, with at least one previous admission (or in one instance, an emergency department attendance) in the previous 2 years. Those living >20 miles from the hospital (n = 4) were excluded as the intervention involved home visiting and attendance at this hospital. One patient with severe mental illness requiring transfer to a psychiatric hospital was also excluded. Addenbrookes Hospital has a catchment area of ~400,000 people, with an estimated 3300 people with Type 1 diabetes. Service
The service included 1.8 full-time equivalent senior diabetes educators with skills in structured education continuous subcutaneous insulin infusion and continuous glucose monitoring systems, 0.1 full-time equivalent diabetes consultants, 0.05 full-time equivalent consultant psychiatrists and 0.4 full-time equivalent psychological wellbeing practitioners. The educators received training in case management,
1362
Preventing diabetic ketoacidosis readmissions D. Simmons et al.
which we defined as the process of planning, coordinating, managing and reviewing the care of an individual. This included liaison with police, housing, benefit and other social services where necessary. The mental health team provided the educators with training, mentoring and supervision in ‘step 2 interventions’ including mental health first aid, such as coping with anxiety, sleep hygiene and positive rewards.
Clinical process
Figure 1 shows the approach used. Essentially, when patients were well enough (generally within 12 h of the hospital diabetic ketoacidosis protocol), the ketoacidosis case manager attended to introduce themselves and to discuss the reasons for the ketoacidosis. A needs assessment was completed as fully as possible. The patients included in the study were often uncooperative with the assessment tool, and some had a history of leaving the hospital on waking (including self-removal of intravenous lines). Once ketones were normalized, and providing there were no acute medical issues requiring hospital stay, the patient went home with follow-up within 48 h at a mutually agreed place (usually home). Patients not contacted during their stay (generally those seen over the weekend), were telephoned after discharge. Frequency of subsequent follow-up was dependent on patient need. Discharge was expedited through the hospital ward team, with the knowledge that a skilled diabetes educator would be in daily contact. Close contact was maintained through 24 h per day telephone availability (either directly by one of the educators or through an ‘out of hours’ advisory service). The patient was invited to attend a weekly dedicated clinic after 1–4 weeks (with the visit supported by the diabetic ketoacidosis case manager) at which he/she: (1) met with a mental health worker who (where agreed) would undertake a more detailed mental health needs assessment and facilitate mental health interventions if needed (e.g. cognitive behavioural therapy for needle phobia, cognitive analytical therapy, antidepressants/ anxiolytics) and (2) met with a diabetes consultant to assess clinical need (e.g. medication/insulin regimen, vomiting management and pain management), identify aspects of care/self-care that could be enhanced, and work with the case manager to develop a plan to improve care/self-care and access mainstream services as soon as possible. A weekly multidisciplinary team meeting took place to provide clinical governance and supervision, including a review of the service and discussion of plans for individual patients. Where the mental health team were unable to assess patients directly, an indirect assessment took place through feedback from the diabetes specialist team. The mental health team also contributed psychological support, professional development (including the gaining of additional mental health assessment and management skills) and supervision for the diabetes case management team.
ª 2015 The Authors. Diabetic Medicine ª 2015 Diabetes UK
Research article
DIABETICMedicine
Pa ent admi ed ↓ Ward contacts Case managers ↓ Case manager visits pa ent on ward when pa ent is communica ng (96%) ↓ Rela onship established, assessment tool completed as much as possible ↓ Contact pa ent post discharge (86%)↔Regular contact (62%) ↓ Home Visit (60%) Implement strategies (n = 50 Type 1 diabetes aŌer exclusions) PaƟent educaƟon (47): Sick day plans and hypo management, injec on site management, DAFNE structured Type 1 diabetes educa on if not already completed, foot care Mental health management (3 already; 16 new; 13 refused): Triage to the wider mental health service interven ons including cogni ve analy cal therapy (16 sessions), Ea ng disorder services, depression management, drug and alcohol services, cogni ve behavioural therapy (6-8 sessions) egfor needle phobia and hypoglycaemia fear Technology (17): CGMS (con nuous glucose monitoring system including alarms) (CSII (con nuous subcutaneous insulin infusion+/-low dose longac ng insulin in case of delivery issues) therapy, new meter and pens. Social intervenƟons (16): Facilita ng benefit acquisi on, housing, food bank use, family support, social care. Support (39-all were given out of hours telephone number): Mo va onal interviewing, home visi ng, 24 hour-7 day/week telephone advice access Clinical management (10): Vomi ng management guideline, pain management, foot care, co-morbidity management
↓↑ A ended Dedicated Clinic (74%) ↓ Discharge from case management if ‘stable’ (27%) (As assessed by the team ieunlikely to be readmi ed)
FIGURE 1 Care process and proportion (%) of processes implemented.
Evaluation
Statistics
Patients were accepted onto the caseload consecutively through the intervention period, September 2012 to June 2013. The intervention stopped on 31 July 2013, when the staff funding ceased, although some support to patients continued unfunded, as it was considered unethical/unsafe to cease all contact immediately. Length of stay and tariff payment data were obtained from the hospital case-mix system for the 2 years preceding the intervention period and for the intervention period itself, and was available for the 8-month period following the intervention (1 August 2013 to 28 February 2014). Clinical information was obtained from the health records (e.g. reason for admission, diabetes duration, age at diagnosis, lipids, HbA1c, blood pressure, BMI, medication, complication) for the same period (where available). The educator completed a structured needs assessment with each patient where possible. Although some patients completed these on first contact, others had several meetings to fully engage trust. Unfortunately, only 16 patients completed even part of the structured questionnaire, and so these data are not shown. Funding received for the service from the local NHS health service commissioning body was £179,820.
Data were analysed using SPSS v19 (IBM, Armonk, NY, USA). Results shown are either mean SD values or frequencies (percentages). Statistical significance was taken at the 5% level and was tested by two-tailed tests. Continuous variables were compared among the three time periods using relatedsamples Wilcoxon signed rank-tests, and HbA1c levels were compared using paired t-tests. Multivariate logistic regression was undertaken to assess which variables were associated with ‘success’; that is, where the patient neither died nor was re-admitted after the intervention period. The evaluation was approved as an audit by Cambridge University Hospitals through the Patient Safety Unit.
ª 2015 The Authors. Diabetic Medicine ª 2015 Diabetes UK
Results Overall, 58 patients were admitted with recurrent diabetic ketoacidosis, hyperglycaemia or severe hypoglycaemia over the 11-month period. One patient had Type 2 diabetes and four patients who had Type 1 diabetes were from out of hospital catchment area. The total proportion of patients with non-Type 2 diabetes and recurrent readmissions over 11 months was therefore estimated to be 53/3300 or 1.5/ 1000 per month. Two patients under primary care had been
1363
DIABETICMedicine
misdiagnosed as having Type 2 diabetes, but in fact had Type 1 diabetes requiring appropriate insulin management. One patient was found to have maturity-onset diabetes of the young type 5, and one had secondary diabetes (post-alcoholassociated pancreatitis). One patient was found ‘dead in bed’ during the 11-month intervention [13], and two patients died during the 8-month post-intervention period, giving annual mortality rates of 1.9 and 5.6% during and after the intervention, respectively. The characteristics of the 50 patients without Type 2 diabetes with active case management are shown in Table 1. Two patients had severe hypoglycaemia, rather than hyperglycaemia. Mental health issues were present in more than half and social issues in a quarter of the patients. The majority had had their diabetes for many years. As many patients had previously defaulted clinic and/or tests (laboratory/retinal screening), metabolic and other data were incomplete. Two (4.9%) had a BMI 75 mmol/mol (9.0%), n/N (%) Mean sd (range) BMIj, kg/m2 Mean SD blood pressurek, mmHg LDL cholesterol, mmol/l
35 14a 25 (50) 18 13 (1–54) 3 (6) 28 (56) 16 (32) 5 (10) 1 (2) 13/11 (28) 32 (64) 2 (4) 21 (42) 6 (12) 16.3/9.3 10 4 1 (2) 99 22 11.2 4.2 43/49 (87.8) 24.8 4.7 (15.9–40.9) 129 21/74 12 3.0 1.0
DAFNE, Dose Adjustment for Normal Eating. Numbers in parentheses are proportion of the sample unless stated otherwise. a 17 (34%)
Research: Care Delivery Effectiveness of a multidisciplinary team approach to the prevention of readmission for acute glycaemic events D. Simmons1, S. Hartnell1, J. Watts1, C. Ward1, K. Davenport1, E. Gunn2 and A. Jenaway2 1 Institute of Metabolic Science, Cambridge University Hospitals NHS Foundation Trust and 2Cambridgeshire and Peterborough NHS Foundation Trust, Cambridge, UK
Accepted 8 April 2015
Abstract Aims To describe the effect of a combined diabetes specialist/mental health team approach to prevent readmissions for acute glycaemic events among patients with diabetes. Methods Consecutive patients with diabetes, readmitted to a single hospital for an acute glycaemic condition, were offered one or more diabetes (including assessment, education, medication, technology use and intensive support) and mental health (including assessment, training and therapies) interventions. The pilot service took place over 11 months, with the preceding 24 months and subsequent 8 months serving as control periods.
Of the 58 patients admitted, 50 had Type 1 diabetes and were from within the hospital catchment area, and were discharged home. Of these, 32 (64%) had a pre-existing mental health issue and 14 (28%) had a complex social situation. In all, 96% of patients were met as an inpatient by a team member, and 94% accepted at least one intervention. The mean SD number of admissions per patient/month dropped from 0.12 0.10 to 0.05 0.10 (P < 0.001) during the intervention, increasing, once the intervention ended, to 0.16 0.36 (P = 0.002). The mean sd length of stay similarly decreased and increased (0.6 0.9 to 0.2 0.7 days; P < 0.001 to 0.006) to 0.6 1.4 days (P = 0.003) per patient/month) across the three periods, as did the mean SD tariff paid per patient/ month (₤258.0 374.0 vs ₤92.1 245.0 vs ₤287.3 563.8; P < 0.001 and P = 0.018, respectively). The mean sd HbA1c level dropped from 99 22 to 92 24 mmol/mol (11.2 4.2% vs 10.6 4.3%; P = 0.014) but did not increase after the intervention [89 26 mmol/mol (10.4 4.5%)]. Results
Conclusions The cost and long-term risks of hospitalization among patients with Type 1 diabetes and recurrent admissions can be reduced by a combined specialist diabetes/mental health team approach.
Diabet. Med. 32, 1361–1367 (2015)
Introduction Both hypoglycaemia and diabetic ketoacidosis, although preventable, remain life-threatening complications of Type 1 diabetes [1]. Historically, patients with recurrent admissions for hypoglycaemia and diabetic ketoacidosis have been termed ‘brittle’, and experience poorer quality of life, more complications and premature death [2–4]. Since the early descriptions of brittle diabetes, new approaches to ketoacidosis/hypoglycaemia management have been introduced to reduce inpatient morbidity/mortality [5], newer insulin formulations/delivery systems have been developed [6] and better patient education for self management of Type 1 diabetes [e.g. the Dose Adjustment For Normal Eating Correspondence to: David Simmons. E-mail: [email protected]
ª 2015 The Authors. Diabetic Medicine ª 2015 Diabetes UK
(DAFNE) programme] has been introduced [7]. These developments have been associated with less hypoglycaemia and ketoacidosis [7,8]. To be effective, however, such developments have to be accessible to patients. A recent study showed that, among African-American people in Atlanta, USA [9], 55% of diabetic ketoacidosis cases were readmissions, with insulin discontinuation the leading precipitating cause, particularly as a result of financial issues. Few studies report the characteristics of those who experience recurrent admissions in settings where these modern approaches to diabetes management are available, and where access to insulin should not be a problem, e.g. the National Health Service (NHS) in the UK. Similarly, there are few reports of intervention studies among adults with recurrent admissions for acute glycaemic events [10], although education, case management and
1361
DIABETICMedicine
What’s new? • Patients with recurrent hospitalization for acute glycaemic events mainly have Type 1 diabetes and usually also have mental health and/or social issues. • Multidisciplinary approaches that combine diabetes, mental health, case management and new technology approaches to assist and support patients with recurrent hospitalization can substantially reduce hospital readmissions, length of stay, cost of admissions and HbA1c levels. • Although some patients with recurrent admissions can achieve independence after intervention, the majority require ongoing support. • Patients with alcohol and/or drug-dependency are particularly challenging. psychological interventions have been recommended to prevent acute events among children and adolescents with recurrent admissions [11,12]. A local audit of 473 patients admitted over 12 months (2010/2011) with ketoacidosis, hyperglycaemia or hypoglycaemia included 168 readmissions (14/month). In the present study, we describe the effectiveness of an 11-month diabetes case management service for adult patients with recurrent hospital attendance, which was designed to reduce hospital bed days attributable to glycaemia-related admissions.
Methods Patients
The study included consecutive patients admitted to Addenbrookes Hospital, Cambridge, with either hyperglycaemia or hypoglycaemia, with at least one previous admission (or in one instance, an emergency department attendance) in the previous 2 years. Those living >20 miles from the hospital (n = 4) were excluded as the intervention involved home visiting and attendance at this hospital. One patient with severe mental illness requiring transfer to a psychiatric hospital was also excluded. Addenbrookes Hospital has a catchment area of ~400,000 people, with an estimated 3300 people with Type 1 diabetes. Service
The service included 1.8 full-time equivalent senior diabetes educators with skills in structured education continuous subcutaneous insulin infusion and continuous glucose monitoring systems, 0.1 full-time equivalent diabetes consultants, 0.05 full-time equivalent consultant psychiatrists and 0.4 full-time equivalent psychological wellbeing practitioners. The educators received training in case management,
1362
Preventing diabetic ketoacidosis readmissions D. Simmons et al.
which we defined as the process of planning, coordinating, managing and reviewing the care of an individual. This included liaison with police, housing, benefit and other social services where necessary. The mental health team provided the educators with training, mentoring and supervision in ‘step 2 interventions’ including mental health first aid, such as coping with anxiety, sleep hygiene and positive rewards.
Clinical process
Figure 1 shows the approach used. Essentially, when patients were well enough (generally within 12 h of the hospital diabetic ketoacidosis protocol), the ketoacidosis case manager attended to introduce themselves and to discuss the reasons for the ketoacidosis. A needs assessment was completed as fully as possible. The patients included in the study were often uncooperative with the assessment tool, and some had a history of leaving the hospital on waking (including self-removal of intravenous lines). Once ketones were normalized, and providing there were no acute medical issues requiring hospital stay, the patient went home with follow-up within 48 h at a mutually agreed place (usually home). Patients not contacted during their stay (generally those seen over the weekend), were telephoned after discharge. Frequency of subsequent follow-up was dependent on patient need. Discharge was expedited through the hospital ward team, with the knowledge that a skilled diabetes educator would be in daily contact. Close contact was maintained through 24 h per day telephone availability (either directly by one of the educators or through an ‘out of hours’ advisory service). The patient was invited to attend a weekly dedicated clinic after 1–4 weeks (with the visit supported by the diabetic ketoacidosis case manager) at which he/she: (1) met with a mental health worker who (where agreed) would undertake a more detailed mental health needs assessment and facilitate mental health interventions if needed (e.g. cognitive behavioural therapy for needle phobia, cognitive analytical therapy, antidepressants/ anxiolytics) and (2) met with a diabetes consultant to assess clinical need (e.g. medication/insulin regimen, vomiting management and pain management), identify aspects of care/self-care that could be enhanced, and work with the case manager to develop a plan to improve care/self-care and access mainstream services as soon as possible. A weekly multidisciplinary team meeting took place to provide clinical governance and supervision, including a review of the service and discussion of plans for individual patients. Where the mental health team were unable to assess patients directly, an indirect assessment took place through feedback from the diabetes specialist team. The mental health team also contributed psychological support, professional development (including the gaining of additional mental health assessment and management skills) and supervision for the diabetes case management team.
ª 2015 The Authors. Diabetic Medicine ª 2015 Diabetes UK
Research article
DIABETICMedicine
Pa ent admi ed ↓ Ward contacts Case managers ↓ Case manager visits pa ent on ward when pa ent is communica ng (96%) ↓ Rela onship established, assessment tool completed as much as possible ↓ Contact pa ent post discharge (86%)↔Regular contact (62%) ↓ Home Visit (60%) Implement strategies (n = 50 Type 1 diabetes aŌer exclusions) PaƟent educaƟon (47): Sick day plans and hypo management, injec on site management, DAFNE structured Type 1 diabetes educa on if not already completed, foot care Mental health management (3 already; 16 new; 13 refused): Triage to the wider mental health service interven ons including cogni ve analy cal therapy (16 sessions), Ea ng disorder services, depression management, drug and alcohol services, cogni ve behavioural therapy (6-8 sessions) egfor needle phobia and hypoglycaemia fear Technology (17): CGMS (con nuous glucose monitoring system including alarms) (CSII (con nuous subcutaneous insulin infusion+/-low dose longac ng insulin in case of delivery issues) therapy, new meter and pens. Social intervenƟons (16): Facilita ng benefit acquisi on, housing, food bank use, family support, social care. Support (39-all were given out of hours telephone number): Mo va onal interviewing, home visi ng, 24 hour-7 day/week telephone advice access Clinical management (10): Vomi ng management guideline, pain management, foot care, co-morbidity management
↓↑ A ended Dedicated Clinic (74%) ↓ Discharge from case management if ‘stable’ (27%) (As assessed by the team ieunlikely to be readmi ed)
FIGURE 1 Care process and proportion (%) of processes implemented.
Evaluation
Statistics
Patients were accepted onto the caseload consecutively through the intervention period, September 2012 to June 2013. The intervention stopped on 31 July 2013, when the staff funding ceased, although some support to patients continued unfunded, as it was considered unethical/unsafe to cease all contact immediately. Length of stay and tariff payment data were obtained from the hospital case-mix system for the 2 years preceding the intervention period and for the intervention period itself, and was available for the 8-month period following the intervention (1 August 2013 to 28 February 2014). Clinical information was obtained from the health records (e.g. reason for admission, diabetes duration, age at diagnosis, lipids, HbA1c, blood pressure, BMI, medication, complication) for the same period (where available). The educator completed a structured needs assessment with each patient where possible. Although some patients completed these on first contact, others had several meetings to fully engage trust. Unfortunately, only 16 patients completed even part of the structured questionnaire, and so these data are not shown. Funding received for the service from the local NHS health service commissioning body was £179,820.
Data were analysed using SPSS v19 (IBM, Armonk, NY, USA). Results shown are either mean SD values or frequencies (percentages). Statistical significance was taken at the 5% level and was tested by two-tailed tests. Continuous variables were compared among the three time periods using relatedsamples Wilcoxon signed rank-tests, and HbA1c levels were compared using paired t-tests. Multivariate logistic regression was undertaken to assess which variables were associated with ‘success’; that is, where the patient neither died nor was re-admitted after the intervention period. The evaluation was approved as an audit by Cambridge University Hospitals through the Patient Safety Unit.
ª 2015 The Authors. Diabetic Medicine ª 2015 Diabetes UK
Results Overall, 58 patients were admitted with recurrent diabetic ketoacidosis, hyperglycaemia or severe hypoglycaemia over the 11-month period. One patient had Type 2 diabetes and four patients who had Type 1 diabetes were from out of hospital catchment area. The total proportion of patients with non-Type 2 diabetes and recurrent readmissions over 11 months was therefore estimated to be 53/3300 or 1.5/ 1000 per month. Two patients under primary care had been
1363
DIABETICMedicine
misdiagnosed as having Type 2 diabetes, but in fact had Type 1 diabetes requiring appropriate insulin management. One patient was found to have maturity-onset diabetes of the young type 5, and one had secondary diabetes (post-alcoholassociated pancreatitis). One patient was found ‘dead in bed’ during the 11-month intervention [13], and two patients died during the 8-month post-intervention period, giving annual mortality rates of 1.9 and 5.6% during and after the intervention, respectively. The characteristics of the 50 patients without Type 2 diabetes with active case management are shown in Table 1. Two patients had severe hypoglycaemia, rather than hyperglycaemia. Mental health issues were present in more than half and social issues in a quarter of the patients. The majority had had their diabetes for many years. As many patients had previously defaulted clinic and/or tests (laboratory/retinal screening), metabolic and other data were incomplete. Two (4.9%) had a BMI 75 mmol/mol (9.0%), n/N (%) Mean sd (range) BMIj, kg/m2 Mean SD blood pressurek, mmHg LDL cholesterol, mmol/l
35 14a 25 (50) 18 13 (1–54) 3 (6) 28 (56) 16 (32) 5 (10) 1 (2) 13/11 (28) 32 (64) 2 (4) 21 (42) 6 (12) 16.3/9.3 10 4 1 (2) 99 22 11.2 4.2 43/49 (87.8) 24.8 4.7 (15.9–40.9) 129 21/74 12 3.0 1.0
DAFNE, Dose Adjustment for Normal Eating. Numbers in parentheses are proportion of the sample unless stated otherwise. a 17 (34%)
