Family Practice Vol. 17, No. 2, 167-172
© Oxford University Press 2000
Quality of life and effectiveness of diabetes care in three different settings in Leuven
a Department of General Practice, KULeuven, Kapucijnenvoer 33 and
b Medisch Centrum voor Huisartsen, Leuven, Belgium.
Van Loon H, Deturck L, Buntinx F, Heyrman J, Degroote L, De Koker K and Vliers J. Quality of life and effectiveness of diabetes care in three different settings in Leuven. Family Practice 2000; 17: 167–172.
Received 10 June 1999; Revised 15 October 1999; Accepted 26 October 1999.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and methodsResultsDiscussionReferences |
|---|
Background. The new diabetes protocol, formulated in Belgium as a consensus between the National Institution of Health Insurance and diabetologists, implicitly assumes the care of type 1 diabetic patients to be more efficient at the specialist level (SP) in hospital, although GPs frequently are involved in diabetes care.
Objectives. A study was carried out in order to highlight differences in diabetes care between three different treatment settings (SP alone, combined SP–GP and GP alone)
Methods. Out of a group of known diabetics, 325 patients were selected according to a stratified cluster sampling technique, in such a way that the three types of diabetes (formerly called type 1, type 2a and type 2b) occurred sufficiently in the three above-defined treatment settings. Outcome data on co-morbidity and diabetes health profile as well as output data on laboratory results were collected for each patient and compared between the different subgroups.
Results. On the basis of a response rate of 47.9%, equally distributed over the different levels, we demonstrated that GPs and SPs score equally low on the different measures and that a better follow-up is indicated in all settings.
Conclusion. Diabetes care in Flanders can be upgraded significantly. There is no evidence that specialists are performing better. Therefore, one could argue for better follow-up of diabetes care in a primary health care setting.
Keywords. Belgium, diabetes care, effectiveness, quality of life.
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and methodsResultsDiscussionReferences |
|---|
Recently1 a new diabetes care protocol has been formulated in Belgium as a consensus between the National Institution of Health Insurance and diabetologists. This protocol stresses the conditions and the responsibilities to be fulfilled by diabetic patients (type 1: insulin-dependent) and their carers (the medical specialist) in order to be eligible for reimbursement of the treatment costs. Implicitly, the care of type 1 diabetic patients has been assumed to be more efficient at the specialistic level (SP) in hospital, although GPs frequently are involved in diabetes care.2 The present study was conducted in order to highlight differences in diabetes care between three different treatment settings (SP alone, combined SP–GP and GP alone).
|
Subjects and methods |
|---|
|
TopAbstractIntroductionSubjects and methodsResultsDiscussionReferences |
|---|
In a small survey among the 57 GPs in the region of Leuven in 1996, 617 diabetic patients were enrolled, of whom 81 could be classified as type 1 (insulin-dependent), 351 as type 2a (adult onset, NIDDM) and 239 as type 2b (adult onset, insulin-treated) (Table 1
). Based on the judgement of the GP, three different treatment settings could be defined: in the first setting, the GP takes care exclusively of the control and follow-up of the diabetic patient (n = 427); in the second setting, caregiving was shared by the SP and the GP (n = 173); and in the third setting, patients were followed exclusively by the SP with regard to their diabetes (n = 71). The aim of the study was to determine whether any difference exists in effectiveness and/or quality of life of the patients in the three different treatment settings.
|
Based on available data in the literature (mean HgbA1c = 7.5, SD = 1.5), the sample size in order to discern differences of a mean HgbA1c of 1% (accepted as clinically relevant) was estimated to be 28 for both groups (e.g. HA versus SP) (
= 0.05, ß = 0.20, one-sided, n1 = n2). Furthermore, to take into account a substantial number of non-responders and to create the possibility for multivariate analysis, a target number of 50 patients in each subgroup (if possible) has been suggested. Out of the 671 patients, a stratified random sample was taken in such a way that, if available, 50 cases in each subgroup should be investigated; this resulted in a total number of 325 patients (Table 1
).
Three kinds of data were collected. With the consent of the GPs, the ‘Diabetes Health Profile’ (DHP) questionnaire, according to Meadows et al.,3 was sent to 325 candidates, of whom 151 responded. This DHP questionnaire is a tool to measure the quality of life of the patients and has been validated by Goddijn et al.4 For these responders, two types of data were added. The first was a questionnaire evaluating the ‘co-morbidity status’, according to Charlson et al.5 The morbidity items were filled out by the GP and are displayed in Table 5. Secondly, the results of recently performed laboratory tests were registered from notes, such as HgbA1c, creatinine and cholesterolaemia.
|
The number of patients in this study is shown in Table 2
. There was an overall response rate of 151/325 (47.9%). An underreporting of the numbers of the combined setting (SP–GP) could be observed in comparison with the others. All calculations and tests were processed with the SAS statistical software package release 6.12 from the SAS Institute.6
|
|
Results |
|---|
|
TopAbstractIntroductionSubjects and methodsResultsDiscussionReferences |
|---|
Descriptive statistics
The study population comprises (Table 2
) 30 patients with type 1 (insulin-dependent), 51 type 2a (adult onset, NIDDM) and 70 type 2b (adult onset, treated with insulin). The diabetes care was provided solely by the GP in 70 cases, a shared care system (SP–GP) was operational in 44 cases, and 37 patients were treated solely by the specialist in a hospital environment.
Table 3 shows that 42.9% of the participants were men and 57.1% women; 18.3% of the patients were younger than 50 years of age, 36.1% ranged between 50 and 69 years, and 45.6% were older than 70 years. The overrepresentation of females in the sample is due solely to the oldest age group (>70 years). Although diabetes is more common in men, there are, in terms of absolute numbers, more diabetic women above 70 years, which is in agreement with the national health survey of 1997.7 The mean time after onset of diabetes was 12.9 years (Table 4). The mean height of the patients was 165.7 cm (SD 10.1), and the mean values for the variables weight, body mass index (BMI) and blood pressure were 73.2 kg, 26.8% and 137.7/80.7 mmHg, respectively. Twenty-five per cent of the patients had a BMI >30%. In the same table, the mean value and the range of recent laboratory results is given for HgbA1c (mean = 7.9%), creatinine (mean = 1.1) and cholesterol (mean = 215). It was of note that the mean HgbA1c value was rather high (mean = 7.9%) while a target value of 6.5% has been defined, with an extreme maximum of 7.5%.5
|
|
Concerning the DHP according to Meadows et al.,2 the items were regrouped into three subscores: 14 items modelling the ‘psychological distress' (PD) component, 12 items assessing the component ‘barriers to activity' (BA), and five items for ‘disinhibited eating' (DE). The PD score varied from 0 to 72, the BA score from 0 to 56 and the DE score from 0 to 90 (Table 4
), where a value of 100 reflects, for each of the components, the presence of a significant problem.
Based on the presence of concomitant diagnoses, in 89 cases one or more co-morbidity factors, as mentioned by Charlson,5 could be identified, from which ‘end organ damage’ was observed most frequently (48.3%) (Table 5). The crude co-morbidity score ranged from 0 to 7 (median = 1), while the co-morbidity–age combined risk score varied from 0 to 11 with a median value of 4.5 (Tables 3 and 6).
|
Univariate comparisons between settings
For each setting, the respective variables have been calculated and are displayed in Table 7
. Based on the values of the different distribution parameters of the laboratory tests, it could be concluded that there was no difference at all between the different treatment settings. Neither the values of HgbA1c, nor cholesterol and creatinine were significantly different (P > 0.05).
|
The same was true for the DHP score. No significant differences between the different settings could be observed (P > 0.05), either for the PD score, the BA score or the DE score.
Nevertheless, a significant relationship was found between the type of diabetes and the PD (P = 0.009) and BA score (P = 0.015), i.e. patients affected by type 2a had lower scores for both DHP scores compared with the other types of diabetes, which can lead to the interpretation that insulin treatment affects the quality of life of the patients in a significant manner.
With regard to the combined age–co-morbidity risk score, a significant difference between the three settings was noted (P = 0.001). However, the direction of the association was different and opposite to that expected because patients with a lower co-morbidity score were those whose treatment was provided by the SP.
Multivariate analysis of the data
Prediction of laboratory results.
In a multifactorial analysis of variance (ANOVA), in which the patient variables were used as covariates, no significant statistical difference between the different settings could be found which could be used to predict the laboratory results; it has to be borne in mind, however, that the sample size was too small for straight-forward conclusions.
HbA1c has been found to be independent of the type of diabetes, the setting, age, gender, creatinine, PD score and DE score (R2 = 0.20). After the progressive reduction of the non-significant independent variables in the final model (P = 0.01), both the BA score (P = 0.03) and the cholesterol values (P = 0.02) were found to be positive and significant (R2 = 0.11).
Prediction of co-morbidity. In the prediction of the combined score of co-morbidity and age following the same approach, only the variable creatinine, apart from age, played a significant, positive role (P = 0.001) (R2 = 0.60). Neither the type of diabetes, nor the setting had any effect.
Prediction of quality of life scores. In the prediction of the quality of life variables such as PD score (psychological distress), a significant effect (R2 =0.71) was discernible for three factors: gender (P = 0.0007), BA score (barriers to activity) (0.0001) and DE score (disinhibited eating) (P = 0.006). Two factors were borderline significant: type of diabetes (P = 0.07) and age (P = 0.07). The BA score could be predicted (R2 = 0.70) by the scores of HbA1c (P = 0.01), PD (P = 0.0001) and DE (P = 0.01). Furthermore, in order to explain the variation in the DE score (R2 = 0.42), the two remaining quality of life scores could be retained, i.e. the PD score (P = 0.02) and the BA score (P = 0.0002). Again, no influence of setting or type of diabetes could be demonstrated (P > 0.05) for any quality of life indicators.
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and methodsResultsDiscussionReferences |
|---|
Based on the results obtained and in terms of the normative values of HgbA1c (<7%),8 it could be concluded that the treatment of diabetes in the Leuven region was somewhat mediocre (Table 4
: median value = 7.8 and Q1 = 6.8). The values were also high for the different types of diabetes and for the different settings. This observation seems to be very important. Recently, Skyler9 argued for adequate glycaemic control, although less aggressive control in very old patients. The arguments were based on epidemiological data and simulation models. A good glycaemic control (HgbA1c <7%) could reduce the cumulative incidence of blindness, end-stage renal disease and lower extremity amputation by 72, 87 and 67%, respectively. On the other hand, good management of diabetes care is not easy and probably only effective in a good comprehensive setting where several disciplines work out a multidisciplinary approach, which can be implemented appropriately and is adaptable to each individual.10 Satisfactory results have already been obtained in Germany,11 Sweden12 and The Netherlands,13 where teamwork (collaboration with nurses, dieticians and GPs) was essential to implement a well-developed strategy with attention to education, diet and physical activity in combination with necessary interventions,14 such as the assessment of cardiovascular risk (including stopping smoking and regular physical exercise), control of blood pressure and annual cholesterol control, daily prophylactic administration of aspirin, annual testing for albuminuria, annual screening for eye disease, inspection of the patient's feet and monitoring of glycaemic control.
Secondly, the generally accepted assumption that diabetes control is more effective if executed by a specialist as compared with a GP is not supported by our data. The laboratory results as well as the health profile (reported by the patients) and the co-morbidity score were not significantly affected by the individual settings. The low response rate (49.7%) made the straigthforward interpretation of these data somewhat difficult. Nevertheless, the underrepresentation of the GP–SP group (possible selection bias) will only affect the overall conclusions in a limited way, since the same observations were made for the different subgroups. Therefore, because no significant differences could be observed, diabetic control by a GP appears to match that by a specialist, which is indirect support for the thesis that no easy and simple approach is available. Even if one argued that the more complicated and difficult cases were already selected by the specialist and therefore no difference could be found between the different settings, this thesis is not supported by our data since the co-morbidity scores were distributed equally in the subgroups, and the combined risk score of co-morbidity and age was distributed in the opposite way to that expected (Table 4). Although this selection bias could not be excluded definitely by cross-sectional data (a longitudinal study with random allocation for setting would be a better design for this purpose), in this data set the more ‘complex’ cases were managed by the GP.
Looking at the quality of life indicators, such as psychological distress (PD), barriers to activity (BA) and disinhibited eating (DE), no differences between the various settings could be found, not even in a multivariate analysis. The low response rate, however, limits the interpretation of the results. On the other hand, there is no clearcut argument in favour of a better performance of either the specialist- or GP-controlled treatment setting. A definite answer cannot be put forward. At least it can be argued that no evidence actually exists (either based on quality or on health reasons) in favour of the government option to promote hospital treatment by a favourable financial incentive.
An intensive follow-up of chronic diseases in general, and of diabetes in particular, organized by the primary health care team,11,12,13 led by the GP in a comprehensive approach and supported by the specialist, is a more logical and probably more effective alternative.
|
References |
|---|
|
TopAbstractIntroductionSubjects and methodsResultsDiscussionReferences |
|---|
1 RIZIV. Verzekeringscomité, Revalidatieovereenkomst inzake Zelfregulatie van Diabetes Mellitus. 1995, 95/216.
2
Griffin S. Diabetes care in general practice: meta-analysis of randomised control trials. Br Med J 1998; 317: 390–396.
3 Meadows K, Steen N, McColl E, Eccles M, Shiels C, Hutchinson A. The Diabetes Health Profile (DHP), a new instrument for assessing the psychosocial of insulin requiring patients, development and psychometric evaluations. Qual Life Res 1996; 5: 242–254.[Web of Science][Medline]
4 Goddijn P, Bilo H, Meadows K, Groenier K, Feskens E, Meyboom de Jong B. The validity and reliability of the Diabetes Health Profile (DHP) in NIDDM patients referred for insulin therapy. Qual Life Res 1996; 5: 433–442.[Web of Science][Medline]
5 Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic co-morbidity in longitudinal studies: development and validation. J Chron Dis 1987, 40: 373–383.[Web of Science][Medline]
6 SAS Institute Inc. SAS Campus Drive, Cary, NC 27513, USA.
7 Demarest J, Tellier V, Van der Heyden J, Schiettecatte E, Taffereau E, Van Oyen H. Health Interview Survey, 1997. Interviewers' Guide. SIP: Brussels Department of Epidemiology, 1996.
8 Van Crombrugge P. Een Interdisciplinaire Consensus over Het Beleid van niet Insuline Dependente Diabetes Mellitus in Vlaanderen. 1997.
9 American Diabetes Association. Standards of medical care for patients with diabetes mellitus. Diabetes Care 1997; 20 (Suppl 1): s5–s13.
10
Skyler J. Glucose control in type 2 diabetes mellitus. Ann Intern Med 1997; 127: 837–839.
11
Berger M, Jörgens V, Flatten G. Health care for persons with non-insulin-dependent diabetes mellitus (The German Experience). Ann Intern Med 1996; 124: 153–155.
12
Falkenberg M. Diabetes care in a rural primary health care district where patient education is given high priority. Metabolic evaluation. Fam Pract 1990; 7: 270–272.
13 de Sonnaville JJJ, Bouma M, Colly LP, Deville W, Wijkel D, Heine RJ. Sustained good glycaemic control in NIDDM patients by implementation of structured care in general practice: 2-year follow-up study. Diabetologia 1997; 40: 1334–1340.[Medline]
14 Kerr CP. Improving outcomes in diabetes: a review of the outpatient care of NIDDM patients. J Fam Pract 1995; 40: 63–75.[Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  R. van Bruggen, K. Gorter, R. Stolk, O. Klungel, and G. Rutten Clinical inertia in general practice: widespread and related to the outcome of diabetes care Fam. Pract., December 1, 2009; 26(6): 428 - 436. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. A Donker, D. M Fleming, F. G Schellevis, and P. Spreeuwenberg Differences in treatment regimes, consultation frequency and referral patterns of diabetes mellitus in general practice in five European countries Fam. Pract., August 1, 2004; 21(4): 364 - 369. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||