Family Practice Advance Access originally published online on November 1, 2006
Family Practice 2007 24(1):20-25; doi:10.1093/fampra/cml055
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Logistic support service improves processes and outcomes of diabetes care in general practice
a Centre for Quality of Care Research, University Medical Centre Nijmegen The Netherlands
b Department of General Practice, Radboud University Nijmegen The Netherlands
c Department of Family Medicine, Maastricht University The Netherlands
d Diabetes Service of the Diagnostic Centre Eindhoven The Netherlands
Correspondence to: Radboud University Nijmegen Medical Centre, Centre for Quality of Care Research (WOK), 117 KWAZO, Drs M. Meulepas, PO Box 9101, 6500 HB NIJMEGEN, The Netherlands. Email: m.meulepas{at}kwazo.umcn.nl
Received 21 January 2006; Accepted 2 October 2006.
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionReferences |
|---|
Background. Guidelines for type 2 diabetes care in general practice are well known and accepted, but the implementation falls short.
Objective. To implement these guidelines by introducing a diabetes support service (DSS) to support the care delivered by the GP.
Methods. A controlled, non-randomised study with delayed intervention in the control group; 78 GPs (n = 51 for the intervention and n = 21 for the control group) in the south of the Netherlands and 613 of their type 2 diabetic patients participated. Data were collected on the frequency, content and results of the check-ups (fasting blood glucose, HbA1c, cholesterol, cholesterol/HDL ratio, triglycerides, creatinine, blood pressure, fundus photography, foot examination, body mass index and smoking status) for 3 years. The year before signing up with the DSS was taken for the pre-measurements and after 2 years of DSS the post-measurements took place. The effect of the DSS was analysed in a mixed model with repeated measurement covariance structure.
Results. At baseline the intervention and control group did not differ in control frequency and outcome (HbA1c). After the intervention the percentage of patients that attended four or more quarterly check-ups (with at least testing of fasting blood glucose or HbA1c) increased from 59 to 78%. In contrast, the frequency of check-ups in the control group remained constant. This effect was significant. The HbA1c remained the same in the intervention group while there was a significant deterioration in the HbA1c in the control group.
Conclusion. Simple logistic support by the DSS proved to have the capacity to implement type 2 diabetes guidelines in general practice.
Keywords. Ambulatory monitoring, diabetes mellitus, general practice, guideline adherence, quality of health care.
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
National and international guidelines for type 2 diabetes care in general practice are especially directed towards strict metabolic control and management of the other risk factors for cardiovascular disease.1–3 However, studies indicate that the guidelines are not adequately followed and many type 2 diabetic patients run a high risk of complications.4,5 There is an enormous variety in care offered in the general practice setting.6 Looking for reasons for this variety, Khunti7 mention 54 possible factors at the level of the practice, the organisation and the patient. A study carried out in 169 practices with a total of 18 642 diabetic patients in the UK, in which the relation between the various factors and the quality of care was investigated, showed that most of the factors did not have a proven effect on the quality of care.8 However, a positive correlation was established between using an active recall system and the quality of diabetes care. Running an active recall system may have a positive effect on diabetes care but it also constitutes a heavy burden on the practice,9 which is why solutions are sometimes sought at a level beyond the single practice.10–12 It appears that when an active recall system is combined with a structured delivery of care from a supporting service, poorly controlled patients reach better metabolic values.13,14 A disadvantage of this approach is that the care is taken away from the responsibility of the GP. In this way the most important task of the GP, to provide integrated care for all patients with regard to all aspects of their health and well-being, is frustrated.15,16 Studies show that good diabetes care can be delivered in the general practice setting as long as the care is well structured.17,18 Moreover, structure in general practice brings about a high level of participation among patients.19,20 Separating the care from the organisation may be the key. Earlier, Hurwitz21 described a successful experiment with a prompting system. Patients were brought in from outpatient clinics. We could not find any examples of centrally organised prompting for coordinating primary care outside the hospital. Therefore, we would like to introduce a diabetes support service (DSS) in primary care that offers logistic support in organising the care but leaves the actual provision of care in the hands of the GP. The effect of this approach on the care process and outcome has, however, not yet been studied. The objective of our study is to determine the effect of the DSS on the implementation of the guidelines by measuring process and outcome indicators of type 2 diabetes care.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Design and study population
The effect of the DSS was investigated in a controlled, non-randomised study with delayed intervention in the control group. In order to make both groups as comparable as possible, GPs on the waiting list for the DSS were eligible for the control group. A pre- and post-assessment was performed with an interval of 2 years.
The study was carried out in the south of the Netherlands among 78 GPs with a total of 613 type 2 diabetic patients. GPs who used the DSS (intervention group, n = 51) were compared with GPs who were not yet using the DSS (control group, n = 27). GPs were eligible if they signed up in 1999 and at the start of the study (end 2001) had at least 10 diabetic patients registered with the DSS (intervention group) or on the waiting list (control group). Furthermore, in the intervention group patients were included at the start of the study when they had been registered with the DSS for a minimum of 1 year and a maximum of 2 years at the start of the study. All patients in the study had had documented diabetes for more than 4 years at the start of the study: Of the 1292, 449 patients in the intervention group versus 164 of the 553 in the control group met the inclusion criteria. Only patients with data available for the whole study period were included in the analysis. Patients who died or moved away from the area during the course of the study were excluded.
The power analysis was based on the assumption that the implementation strategy could lower the HbA1c of type 2 diabetic patients by 0.5 (SD 1.5). We corrected for known unbalanced groups (intervention group 70% and control group 30%). The clustering of patients per GP was also taken into account. With alpha = 5%, power = 80%, ICC = 0.05, 75 GPs with a total number of 425 patients were needed (298 intervention and 127 control patients). With an expected dropout of 10%, the total number of patients needed was 468.
Intervention
The intervention consisted of logistic support for the GP by a DSS. The DSS called up patients for laboratory testing, foot examination, fundus photography and appointments with the dietician and the diabetes nurse; the patient was asked to make an appointment at the surgery to discuss the results.
The DSS did not provide any patient care itself; there was no contact with patients except for calling them up for (repeated) three-monthly and annual blood testing. For this purpose the DSS worked together with a laboratory that offered blood testing centrally, but also decentralised at general practice surgeries, health care centres, etc.
The results of the requested tests were sent directly from the laboratory to the GP. This was the same procedure as for the GPs in the control group who could request regular testing by the laboratory. The actions taken by the GPs after receiving the results were not part of the intervention.
Variables and instruments
Data were collected from all the patients in the research group on the frequency, content and results of the check-ups (fasting blood glucose, HbA1c, cholesterol, cholesterol/HDL ratio, triglycerides, creatinine, blood pressure, fundus photography, foot examination, body mass index and smoking status). In addition, a number of general characteristics of the GPs were noted: age, level of urbanisation of the place where the surgery was situated, size of the practice, percentage of employment of the GP, and number of known diabetic patients. Of the patients, age and sex were noted as general characteristics.
The data were obtained from the databases of the DSS and the associated laboratory. Missing data—for instance if the GP does the quarterly monitoring of fasting blood glucose in his own practice and the result is therefore not in the laboratory database—were collected from the practice.
Analysis
Differences in the frequency of tests and the test results between intervention and control groups were assessed with mixed models with repeated measures (Proc mixed procedure SAS V8.2).22 Test results in as far as they were categorised (within the target value or not) and differences in whether or not the patient underwent four check-ups a year were assessed with a mixed logistic model repeated measures (Glimmix procedure SAS V8.2) over the year prior to registering with the DSS and the 2 year after registration.
All analyses were corrected for the random/cluster effect caused by patient and GP.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Response rate and possible selective loss
In total 78 GPs (out of 82) met the inclusion criteria and they all took part. Of their diabetic patients, 613 met the inclusion criteria of whom 497 patients (80%) gave consent to use of their details for study purposes; 15 patients (2%) of the latter group moved away from the area or died. Finally, data from 482 patients were analysed (Overview 1) (Fig. 1). Due to the combination of diabetes duration and date of intake with the DSS half of the patients did not meet the inclusion criteria. Comparison of metabolic values at the time of registering with the DSS, however, shows that there were no significant differences between patients who met the inclusion criteria and those who did not. The GPs in the control group were comparable with those in the intervention group as regards age, proportion of rural and urban practices, population size and number of known diabetic patients. The patients in the control group did not differ in age and sex from the intervention group (Table 1).
|
|
Frequency and content of check-ups
In the year prior to registration with the DSS, 59% of the patients from the intervention group attended four or more quarterly controls (with at least testing of fasting blood glucose or HbA1c) versus 49% of the patients in the control group. The difference between intervention and control group at baseline was considerable but not significant. After the intervention this percentage increased to 78% in the 2 year, but remained constant in the control group (P < 0.0001).
Table 2 also shows the tests that were carried out at least once a year, generally during the annual check-up. A percentage of 100% means that a certain test was not only ordered by the GP but that all the patients actually attended.
|
Before the intervention patients underwent on average four tests, as did those in the control group. These were mainly blood tests and blood pressure measurement. For the patients in the control group this did not change in the following years, while the intervention patients had undergone an average of eight of the nine tests on offer after 2 years. Foot examinations, fundus photography, and questioning about the smoking status gained most by the DSS.
Outcome of care
The patients in the intervention group had a lower mean fasting blood glucose in the year prior to registration with the DSS than the control patients, although this difference was not significant (Table 3). The HbA1c was 0.2% lower in the intervention group. Two years later, there was a significant deterioration in the HbA1c in the control group and no difference in the HbA1c in the intervention group. The difference in change between the two groups was 0.7%. The mean cholesterol values did not significantly differ within the groups and between the groups, but there was a shift in the percentage of patients with a cholesterol >5 mmol. That was initially higher in the intervention group (63 versus 56%), but had dropped significantly after 2 years (51 versus 66%).
|
The mean systolic blood pressure in intervention patients was significantly lower than in the patients in the control group. The diastolic blood pressure did not significantly differ.
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Simple logistic support by a DSS without taking over patient care improved adherence to general practice guidelines for diabetes care. The recall system produced an increase in the number of patients who, in accordance with the guidelines, underwent four check-ups a year. Also the content of the check-ups was more in line with the guidelines than for patients not supported by the DSS. Modest but significant improvement or less worsening was seen in mean levels of HbA1c, systolic blood pressure, cholesterol and triglycerides, but not in diastolic blood pressure and cholesterol/HDL ratio. The rise in mean HbA1c levels in the control group was similar to the UKPDS trends in HbA1c levels.23
There are some methodological considerations. Mixed models repeated measures are appropriate for nested analyses, but since we did not use a randomised design there is a potential bias in the selection of the candidates that is not corrected by analysis. In fact the mean entry value of the primary outcome measure HbA1c in the control group is high (7.4). Goudswaard et al.24 found a mean HbA1c of 7.1 in a population of 1641 patients. So the intervention group resembles the mean population of Dutch diabetic patients in primary health. In that group the DSS appeared to be capable of preventing deterioration in metabolic levels. Secondly, the inclusion criteria resulted in half of the patients not being included in the study. This loss was not selective for the metabolic values that were checked at baseline and did not differ significantly. We therefore assume that the study population provides a representative sample of the group of patients GPs ask the DSS to call up. Thirdly, the fact that the control group were on the waiting list could disturb the attribution of the effect. Classical threats are selection maturation, differential statistical regression and local history.25 However, these treats do not seem very likely, because the waiting list problem was created by a totally unexpected national funding policy by health insurance companies that was corrected for after our intervention period.
After 2 years of support by the DSS the majority of tests are carried out in 100% of the patients which demonstrates the high level of willingness among patients to respond to the call up by a DSS. This is in line with the high level of participation among diabetic patients who are offered well-structured care from the general practice setting that is reported in the literature.18,19 On the other hand, improvement in the blood glucose and blood pressure control of patients was less convincing than interventions described in the literature that focused on both the organisation and the GP.14,15 The UKPDS showed us that intensive blood pressure control in diabetic patients might be even more important than blood glucose control,26 but with the support of the DSS target levels of the latest guidelines (140/85) could not be reached. Logistic support appears to be important for the improvement of the healthcare processes, but it should be possible to achieve a greater improvement in patient outcome. Further research is needed to investigate whether other forms of support in addition to the logistic support would produce more effect on the patient outcome. Options could include support in the surgery in discussing test results with the patient, as well as the consequences for treatment.
|
Notes |
|---|
Meulepas MA, Braspenning JCC, de Grauw WJ, Lucas AEM, Harms L, Akkermans RP and Grol RPTM. Logistic support service improves processes and outcomes of diabetes care in general practice. Family Practice 2007; 24: 20–25.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
1 Rutten GEHM, Verhoeven S, Heine RJ, et al. (1999) Dutch College of General Practitioners guidelines for M01 Diabetes mellitus type 2 (first review) [In Dutch]. Huisarts Wet 42:67–84.
2 The Netherlands Diabetes Federation. (2000) Guidelines and advice for good diabetes care [in Dutch]. Edition.
3 UK Prospective Diabetes Study Group. (1998) Intensive blood-glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes (UKPDS 33). Lancet 352:837–853.[CrossRef][Web of Science][Medline]
4 Beckles GL, Engelgau MM, Venkat Narayan KM, Herman WH, Aubert RE, Wiliamson DF. (1998) Population-based assessment of the level of care among adults with diabetes in the US. Diabetes Care 21:1432–1438.[Abstract]
5 Konings GP, Wijkel D, Rutten GE. (1995) Is it possible to work according to Dutch College of General Practitioners guidelines for diabetes mellitus type II? [in Dutch]. Huisarts Wet 38:10–14.
6 Khunti K, Ganguli S, Lowy A. (2001) Inequalities in provision of systematic care for patients with diabetes. Fam Pract 18:27–32.
7 Khunti K. (1999) Use of multiple methods to determine factors affecting quality of care of patients with diabetes. Fam Pract 16:489–494.
8 Khunti K, Ganguli S, Baker R, Lowy A. (2001) Features of primary care associated with variations in process and outcome of care of people with diabetes. Br J Gen Pract 51:356–360.[Web of Science][Medline]
9 Koch JLM. (1992) Efficacy of education and systematic general practitioner control of diabetes patients: a study in general practices. Nijmegen University [dissertation, in Dutch].
10 Bilo HJG, van Nunen F, van Ballegooie E, Meyboom-de Jong B, Ubink-Veltmaat LJ. Transmural forms of care of diabetes mellitus. A survey of the situation in the year 2000 Isala series 29.
11 De Sonnaville J. (1998) Structured care for patients with diabetes mellitus type 2 in general practice [dissertation]. Amsterdam University.
12 Whitty P, Eccles MP, Hawthorne G, et al. (2004) Improving services for people with diabetes: lessons from setting up the DREAM trial. Practical Diabetes International 21:323–328.
13 De Sonnaville JJ, Bouma M, Colly LP, Devillé W, Wijkel D, Heine RJ. (1997) Sustained good glycaemic control in NIDDM patients by implementation of structured care in general practice: 2 years follow-up study. Diabetologia 40:1334–1340.[CrossRef][Medline]
14 Groeneveld Y. (2002) Introduction of structured diabetes care in general practice [dissertation]. Leiden University.
15 Overland J, Yue DK, Mira M. (2001) Continuity of care in diabetes: to whom does it matter? Diabetes Res Clin Pract 52:55–61.[CrossRef][Web of Science][Medline]
16 Hanninen J, Takala J, Keinanen-Kiukaanniemi S. (2001) Good continuity of care may improve quality of life in type 2 diabetes. Diabetes Res Clin Pract 51:21–27.[CrossRef][Web of Science][Medline]
17 Griffin S. (1998) Diabetes care in general practice: meta-analysis of randomized control trials. BMJ 317:390–396.
18 McCullough DK, Price MJ, Hindmarch M, Wagner EH. (1998) A population-based approach to diabetes management in a primary care setting: early results and lessons learned. Eff Clin Pract 12–22.
19 Freriks JP. (2000) Diabetes services, always a challenge [in Dutch]. TvHG 1:14–19.
20 De Keijzer K. (2000) Volume growth in first-line diabetes services [dissertation, in Dutch]. Utrecht University.
21 Hurwitz B, Goodman C, Yudkin J. (1993) Prompting the clinical care of non-insulin dependent (type II) diabetic patients in an inner city area: one model of community care. BMJ 306:624–630.
22 Little RC, Pendergast J, Natarajan R. (2000) Modelling covariance structure in the analysis of repeated measures data [review]. Stat Med 19:1793–1819.[CrossRef][Web of Science][Medline]
23 Anonymous. (1998) Intensive blood-glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes (UKPDS 33). UK Prospective Diabetes Study (UKPDS) Group. Lancet 352:837–853.[CrossRef][Web of Science][Medline]
24 Goudswaard AN, Stolk RP, Zuithoff P, Rutten GE. (2004) Patient characteristics do not predict poor glycaemic control in type 2 diabetes patients treated in primary care. Eur J Epidemiol 19:541–545.[CrossRef][Web of Science][Medline]
25 Cook TD and Campbell DT. (1979) Quasi-Experimentation, Design and Analysis Issues for Field Settings(Houghton Mifflin Company, Boston).
26 Anonymous. (1998) Tight blood pressure control and risk of macrovascular and microvascular complications in type 2 diabetes: UKPDS 38. UK Prospective Diabetes Study Group. BMJ 317:703–713.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  M A Meulepas, J C C Braspenning, W J de Grauw, A E M Lucas, D Wijkel, and R P T M Grol Patient-oriented intervention in addition to centrally organised checkups improves diabetic patient outcome in primary care Qual. Saf. Health Care, October 1, 2008; 17(5): 324 - 328. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. F. Burden Improving care for patients with diabetes: the role of simple reminders Fam. Pract., February 1, 2007; 24(1): 1 - 2. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||