- Study protocol
- Open Access
- Open Peer Review
This article has Open Peer Review reports available.
Towards a more efficient diabetes control in primary care: six-monthly monitoring compared with three-monthly monitoring in type 2 diabetes - The EFFIMODI trial. Design of a randomised controlled patient-preference equivalence trial in primary care
© Wermeling et al; licensee BioMed Central Ltd. 2010
Received: 22 March 2010
Accepted: 11 May 2010
Published: 11 May 2010
Scientific evidence for the frequency of monitoring of type 2 diabetes patients is lacking. If three-monthly control in general practice could be reduced to six-monthly control in some patients, this would on the one hand reduce the use of medical services including involvement of practice nurses, and thus reduce costs, and on the other hand alleviate the burden of people with type 2 diabetes. The goal of this study is to make primary diabetes care as efficient as possible for patients and health care providers. Therefore, we want to determine whether six-monthly monitoring of well-controlled type 2 diabetes patients in primary care leads to equivalent cardiometabolic control compared to the generally recommended three-monthly monitoring.
Methods and design
The study is a randomised controlled patient-preference equivalence trial. Participants are asked if they prefer three-monthly (usual care) or six-monthly diabetes monitoring. If they do not have a preference, they are randomised to a three-monthly or six-monthly monitoring group. Patients are eligible for the study if they are between 40 and 80 years old, diagnosed with type 2 diabetes more than one year ago, treated by a general practitioner, not on insulin treatment, and with HbA1c ≤7.5%, systolic blood pressure ≤145 mmHg and total cholesterol ≤5.2 mmol/l. The intervention group (six-monthly monitoring) will receive the same treatment with the same treatment targets as the control group (three-monthly monitoring). The intervention period will last one and a half year. After the intervention, the three-monthly and six-monthly monitoring groups are compared on equivalence of cardiometabolic control. Secondary outcome measures are HbA1c, blood pressure, cholesterol level, Body Mass Index, smoking behaviour, physical activity, loss of work due to illness, health status, diabetes-specific distress, satisfaction with treatment and adherence to medications. We will use intention-to-treat analysis with repeated measures. For outcomes that have only baseline and final measurements, we will use ANCOVA. Depending on the results, a cost-minimisation analysis or an incremental cost-effectiveness analysis will be done.
This study will provide valuable information on the most efficient control frequency of well-controlled type 2 diabetes patients in primary care.
Current Controlled Trials ISRCTN93201802
At the end of 2007 more than 660.000 people were diagnosed with type 2 diabetes in the Netherlands . The number of type 2 diabetes patients is still increasing , and also their use of health care facilities. Furthermore, the overall workload of general practitioners is increasing . The current guideline on type 2 diabetes in primary care in the Netherlands advises to monitor type 2 diabetes patients four times a year , but this advice is not evidence-based. Three quarterly controls are done by the practice nurse and the general practitioner is advised to perform the annual check-up. Comparing 15 diabetes guidelines in 13 countries, the advised frequency of monitoring HbA1c ranged from one to four times a year and monitoring blood pressure ranged from one to six times a year . It is obvious that the workload for healthcare professionals will differ significantly, depending on the guideline that is followed.
A retrospective, observational study in Spain demonstrated that the improvement in glycaemic control over time in patients with type 2 diabetes in general practice was not related to the number of visits to the general practitioner, but to changes in treatment . More evidence on the desired frequency of type 2 diabetes controls in general practice is lacking.
A randomised equivalence trial in Canada compared blood pressure control, adherence to treatment and patient satisfaction in patients with treated hypertension followed up by their family physicians every three or six months. Patients with follow-up every three months achieved the same levels of blood pressure control, adherence to treatment and patient satisfaction compared to patients with follow-up every six months .
If, in accordance with the hypertension example, the contact time with the diabetes team in well-controlled type 2 diabetes patients could also be reduced up to 50% without deteriorating their quality of care, this could reduce the patient burden as well as induce savings on direct medical costs and relieve the workload of practice nurses.
Therefore we designed the EFFIcient MOnitoring of DIabetes (EFFIMODI) study, aiming to make primary diabetes care as efficient as possible for patients and health care providers. We hypothesise that six-monthly monitoring of well-controlled patients with type 2 diabetes in primary care results in equivalent cardiometabolic control as the currently recommended three-monthly monitoring, with less costs.
Methods and design
We chose to conduct a patient-preference trial because of two reasons . First, it gives us the possibility to compare the relationship between patients' preference for a frequency of diabetes control and the study outcomes. Second, more people will participate in the study, as people can choose not to be randomised; they can be included in the so-called 'observational' study arm of their choice. Doing so, we will have information about people who are not randomised and this will help in generalising the results. In conducting a patient-preference trial we will avoid selection and probably also drop-out after randomisation.
We chose to conduct an equivalence trial, because we want to assess whether six-monthly monitoring results in equal cardiometabolic control compared to the current frequency of control . Since we expect that six-monthly monitoring will not give better cardiometabolic control than three-monthly monitoring, we did not choose a non-inferiority trial. The intervention period will last one and a half year, so patients will be seen either seven or four times during the intervention period.
The Medical Research Ethics Committee of the University Medical Center Utrecht has approved the study protocol (Protocol number: 08-453).
Patients are eligible for the study if they are between 40 and 80 years old, diagnosed with type 2 diabetes for more than a year, treated by their general practitioner, not on insulin treatment and overall well-controlled, defined as having HbA1c ≤7.5% and systolic blood pressure ≤145 mmHg and total cholesterol ≤5.2 mmol/l. These HbA1c, systolic blood pressure and total cholesterol values are a little higher than the Dutch target values (HbA1c ≤7.0%, systolic blood pressure ≤140 mmHg and total cholesterol ≤4.5 mmol/l ). We decided to choose more liberal values to create a larger target population to be randomised. According to the stricter target values only 18.9% of all type 2 diabetes patients fulfilled these values . Because of the minimal difference with the target values, we assume that 20% of the type 2 diabetes patients will fulfil the inclusion criteria.
Recruitment of practices and patients
We will approach the boards of several health care groups to recruit general practitioners. These health care groups have a central database in which all type 2 diabetes patients are recorded. In this database all determinants that are needed for the selection of study patients are included (see study population for the inclusion criteria). The health care groups will ask all their affiliated general practitioners to participate in the study. If a general practitioner wants to participate he will obtain the selection of patients, according to the inclusion criteria, from the health care group. The general practitioner sends an information letter as well as an informed consent form to the selected patients.
Patients who want to participate have to fill in the informed consent form. The participants are asked whether they strongly prefer three-monthly (current care) or six-monthly diabetes control or whether they have no preference. Patients with a preference will enter the 'observational arm'. If participants do not have a strong preference for the frequency of care, they will be randomised.
Randomisation and blinding
Participants without a preference for either three-monthly or six-monthly monitoring are randomised into one of the two randomised study arms in a 1:1 ratio: a control group that will receive current care, comprising diabetes control once per three months, and an intervention group that will receive diabetes control once per six months. In both randomised groups the extensive annual check-up will be done by the general practitioner. The treatment targets, therapeutic algorithms and life style advices will remain unchanged and do not differ between the intervention and control groups.
Randomisation is generated at the patient level by a computerised random-number generator at the Julius Center. Participants are randomised before baseline data are collected. Since this is a pragmatic trial, it is not necessary to blind participants and general practitioners for the treatment allocation. However, the laboratory technicians who measure HbA1c and cholesterol are not aware of study participation of the patients.
The primary outcome measure is the percentage of people that remains under good cardiometabolic control, defined as having HbA1c ≤7.5% and systolic blood pressure ≤145 mmHg and total cholesterol ≤5.2 mmol/l. Secondary outcome measures are HbA1c, blood pressure, cholesterol, Body Mass Index, fasting blood glucose, smoking behaviour, physical activity, loss of work due to illness, health status, diabetes-specific distress, satisfaction with treatment and adherence to medications. These outcomes are collected either from the medical records or from a patient questionnaire.
Biomarkers (HbA1c, cholesterol, fasting blood glucose) and anthropometric variables (blood pressure, Body Mass Index) will be collected from the general practitioners' computerised information system. Information on the most recently measurements of the biomarkers and anthropometric variables before the study, medical history, medication use before the study and all measurements performed during the first EFFIMODI visit will be collected just after the start of the study. All measurements during the follow-up period will be collected after the end of the follow-up period. The same applies to the number of diabetes and non-diabetes related visits to the general practitioner and (differences in) medication use during the study.
The participants will be asked to complete an extensive questionnaire before (t = 0) and after (t = 18 months) the intervention period. This questionnaire comprises general background information on age, gender, ethnicity, education, smoking, physical activity, occupation and loss of work due to illness, health status, diabetes-specific distress and satisfaction with diabetes treatment.
Current smoking is measured as the number of cigarettes per day or the number of cigars per week. Smoking in the past is measured the same way. Also the number of pack years is recorded. Physical activity is measured with the Short Questionnaire to Assess Health-enhancing physical activity (SQUASH) . The SQUASH is a reliable and valid questionnaire to measure the level of physical activity in an adult population. The questionnaire was designed to give an indication of the habitual activity level. Information on light (range: 2-4 Metabolic Equivalent of Task (MET)), moderate (range: 4-6.5 MET) and vigorous (>6.5 MET) intensity physical activities will be obtained. Physical activity will be expressed in minutes per week and in a total activity score. The total activity score will be calculated by multiplying the minutes per week by the actual MET score of the specific activity (MET/min/week).
Occupation and loss of work due to illness are measured with the Short Form Health and Labour Questionnaire (SF-HLQ) . The SF-HLQ consists of three parts: absenteeism from paid work, production losses without absenteeism from paid work and hindrance in the performance of paid and unpaid work. Data derived with this questionnaire will be used to calculate costs of productivity losses, should incremental cost-effectiveness be merited at closure of the trial (see economic evaluation).
Two questionnaires are used to measure health status: Short-Form 36 (SF-36)  and EQ-5D . The SF-36 generates a profile of scores on eight dimensions of health. These dimensions are: (1) physical functioning; (2) limitations due to physical difficulties (physical role functioning); (3) bodily pain; (4) social functioning; (5) mental health; (6) limitations due to emotional difficulties (emotional role functioning); (7) vitality; and (8) general health perception. For all eight dimensions an average score for all items in the scale is calculated, with a range from 0 (least favourable health state) to 100 (most favourable health state). Two summary scales for mental and physical functioning can be calculated as well. The SF-36 is validated in the Dutch population .
The EQ-5D is a generic questionnaire, consisting of a Visual Analogue Scale (EQ-5D VAS) and a classification system (EQ-5D Profile) . The EQ-5D Profile covers five domains of health (mobility, self-care, usual activities, pain/discomfort and anxiety/depression), each with three levels of functioning: level 1, no problems; level 2, some problems; level 3, severe problems. The EQ-5D VAS is a graded, vertical line, anchored at 0 (worst imaginable health state) and 100 (best imaginable health state). The patient is asked to mark a point on the EQ-5D VAS that best reflects his/her actual health state.
To measure diabetes-specific distress, the Problem Areas In Diabetes (PAID) questionnaire is used . This is a widely recognised measure of diabetes distress, assessing the general emotional burden of diabetes and distress related to treatment, food choices and social support. The 20 items are scored on a 5-point scale yielding a sum score (0-80), with higher scores representing higher distress. The Dutch PAID scale has good convergent and discriminate validity and high internal consistency .
To measure satisfaction with diabetes treatment, the Diabetes Treatment Satisfaction Questionnaire (DTSQ) is used . The DTSQ measures satisfaction with treatment regimen (six items), perceived frequency of hyperglycaemia (one item) and perceived frequency of hypoglycaemia (one item) over the past few weeks. The treatment satisfaction score can range from 0 (very dissatisfied) to 36 (very satisfied).
To be able to calculate direct health care costs, data on health care use are needed. Data concerning consultations beyond the planned monitoring consultations, medication use and referral rates to other health care professionals will be collected from the general practitioners' computerised Information System. Should this trial provide evidence that three-monthly monitoring results in better outcomes than six-monthly monitoring, an incremental cost-effectiveness analysis becomes warranted, that quantifies the additional cost related to the additional health effects. For such an economic evaluation we will use data on health care use as recorded in the information system of the general practitioners, and data recorded with the SF-HLQ and the EQ5D (see above). The EQ5D is of special importance, as utilities, and consequently, quality adjusted life years (QALYs) can be elicited using this generic questionnaire.
Sample size calculation
The sample size is calculated on the assumption of equivalence of cardiometabolic control. Therefore, we used the formula from Jones et al. : n = (2p(100-p)*(Z(1-α)+Z(1-β))2)/δ2, where p is the overall percentage of successes to be expected if the treatments are equivalent, δ indicates the range of equivalence for the difference in percentage success rates, α is the probability of type I error and β the probability of type II error.
We assume equivalence if the two-sided 95% confidence interval (α = 0.05; Z(1-α) = 1.96) for the difference in cardiometabolic control between the two intervention groups is completely in the range from -5 to 5% (δ = 5). With a supposed overall percentage of success of 95% and a power of 90% (β = 0.1; Z(1-β) = 1.28), we need a sample size of 399 people per randomised group. Based on a small survey that was performed by the Julius Center, we assume that ~50% of the people have no preference and thus will be randomised. About the other half of the people is assumed to have a strong preference for the frequency of monitoring. Therefore, we need to include 1596 patients: 798 in the randomised arms and 798 in the observational arms.
As said earlier in the study population section, we assume that 20% of the type 2 diabetes patients will fulfil the inclusion criteria. With an average of 80 type 2 diabetes patients per practice, we expect that sixteen patients in each general practice will meet the eligibility criteria. Taking into account a response rate of 60%, ten patients in each practice will be willing to participate in the study: five in the observational arm and five in the randomised arm, so at least 160 general practitioners will have to be recruited. As people may drop out during the study, we will recruit 1800 patients allowing an 11% drop-out; therefore we need 180 general practitioners. Figure 1 shows the participant flowchart with the expected numbers.
We will use repeated measures analysis for all recorded measurements of HbA1c, blood pressure, cholesterol and fasting blood glucose to optimally use all available data. Data from the questionnaires at the start and at the end of the intervention period will be analysed using ANCOVA. Data will be analysed according to intention-to-treat. For handling missing data we will use multiple imputation . After the intervention, the randomised three-monthly and six-monthly monitoring groups are compared on equivalence of cardiometabolic control.
Should this study demonstrate equivalent outcomes with different frequencies of diabetes monitoring, the evident cost-reduction of less control visits merit the conclusion that six-monthly monitoring is the approach of first choice and a cost-minimisation analysis will be done. This is likely to provide sufficient evidence to change the frequency of monitoring in well-controlled type 2 diabetes patients in general practice.
However, should the three-monthly follow-up scheme result in a better cardiometabolic control, a cost-effectiveness analysis will be performed. If better outcomes can be realised with a higher control frequency at an associated higher cost to society, or vice versa, (somewhat) worse patient outcomes at a lower cost, the balance between costs and outcomes is of interest. Differences in the number of QALYs between groups during the study period will be assessed. The difference in treatment effect will be calculated as follows: for each group the difference between baseline and final measurements in percentage of patients with good cardiometabolic control will be determined. The absolute difference between measurements pre- and post-intervention will be taken as the intervention effect. Accordingly, differences in QALYs between groups will be calculated. The 'incremental cost-effectiveness ratio' (ICER) will be expressed as cost differences between groups divided by differences in treatment effects between groups. Confidence intervals will be determined using bootstrapping . A 'cost-effectiveness acceptability curve' (CEAC) will also be drawn using the bootstrap sample. Cost-utility estimates will be derived accordingly, using QALY differences between groups as outcome measure.
If the results of this study will show that equivalent cardiometabolic control is achieved following six-monthly diabetes control in a sub-sample of generally well-controlled people with type 2 diabetes in general practice as compared to the usual three-monthly diabetes control, their diabetes control frequency can be reduced. This will reduce the use of medical services and direct health care costs, alleviate the burden of a substantial part of the people with type 2 diabetes and relieve the workload of diabetes nurses. If three-monthly monitoring turns out to result in a better regulation of diabetes, a cost-effectiveness analysis is necessary to estimate whether the higher costs of three-monthly follow-up balance the better patient outcomes.
The results of this study will provide valuable information for health care professionals and policy makers on cost-effectiveness of diabetes monitoring. In the case of proven cost-effectiveness, we will recommend implementing a lower control frequency for well-controlled type 2 diabetes patients.
This study is funded by ZonMw, the Netherlands organisation for health research and development (no: 80-82310-98-09058).
- Baan CA, Poos MJJC: Hoe vaak komt diabetes mellitus voor en hoeveel mensen sterven eraan? [in Dutch]. Volksgezondheid Toekomst Verkenning, Nationaal Kompas Volksgezondheid. 2009, Bilthoven, RIVMGoogle Scholar
- van den Berg MJ, Kolthof ED, de Bakker DH, Zee van der J: Tweede Nationale Studie naar ziekten en verrichtingen in de huisartspraktijk: de werkbelasting van huisartsen [in Dutch]. 2004, Utrecht: NIVELGoogle Scholar
- Rutten GEHM, De Grauw WJC, Nijpels G, Goudswaard AN, Uitewaal PJM, van der Does FEE, et al: NHG-standaard Diabetes mellitus type 2 [article in Dutch]. Huisarts Wet. 2006, 49: 137-152.Google Scholar
- Burgers JS, Bailey JV, Klazinga NS, van der Bij AK, Grol R, Feder G: Inside guidelines: comparative analysis of recommendations and evidence in diabetes guidelines from 13 countries. Diabetes Care. 2002, 25: 1933-1939. 10.2337/diacare.25.11.1933.View ArticlePubMedGoogle Scholar
- Ortiz Tobarra MT, Orozco Beltrán D, Gil Guillén V, Terol Moltó C: [Frequency of attendance and degree of control of type-2 diabetic patients]. Aten Primaria. 2008, 40: 139-144. 10.1157/13116629.View ArticlePubMedGoogle Scholar
- Birtwhistle RV, Godwin MS, Delva MD, Casson RI, Lam M, MacDonald SE, Seguin R, Ruhland L: Randomised equivalence trial comparing three month and six month follow up of patients with hypertension by family practitioners. BMJ. 2004, 328: 204-10.1136/bmj.37967.374063.EE.View ArticlePubMedPubMed CentralGoogle Scholar
- Torgerson DJ, Sibbald B: Understanding controlled trials. What is a patient preference trial?. BMJ. 1998, 316: 360-View ArticlePubMedPubMed CentralGoogle Scholar
- Jones B, Jarvis P, Lewis JA, Ebbutt AF: Trials to assess equivalence: the importance of rigorous methods. BMJ. 1996, 313: 36-39.View ArticlePubMedPubMed CentralGoogle Scholar
- Cleveringa FG, Gorter KJ, van den Donk M, Rutten GE: Combined task delegation, computerized decision support, and feedback improve cardiovascular risk for type 2 diabetic patients: a cluster randomized trial in primary care. Diabetes Care. 2008, 31: 2273-2275. 10.2337/dc08-0312.View ArticlePubMedPubMed CentralGoogle Scholar
- Wendel-Vos GC, Schuit AJ, Saris WH, Kromhout D: Reproducibility and relative validity of the short questionnaire to assess health-enhancing physical activity. J Clin Epidemiol. 2003, 56: 1163-1169. 10.1016/S0895-4356(03)00220-8.View ArticlePubMedGoogle Scholar
- Hakkaart-van Roijen L, Bouwmans CAM: Manual Short Form-Health and Labour Questionnaire (SF-HLQ). 2007, RotterdamGoogle Scholar
- Ware JE, Sherbourne CD: The MOS 36-item short-form health survey (SF-36). I. Conceptual framework and item selection. Med Care. 1992, 30: 473-483. 10.1097/00005650-199206000-00002.View ArticlePubMedGoogle Scholar
- Brooks R: EuroQol: the current state of play. Health Policy. 1996, 37: 53-72. 10.1016/0168-8510(96)00822-6.View ArticlePubMedGoogle Scholar
- van der Zee KL, Sanderman R: Het Meten van de algemene gezondheidstoestand met de RAND-36. Een handleiding [in Dutch]. 1993, Groningen: Northern Center for Health Care ResearchGoogle Scholar
- Polonsky WH, Anderson BJ, Lohrer PA, Welch G, Jacobson AM, Aponte JE, Schwartz CE: Assessment of diabetes-related distress. Diabetes Care. 1995, 18: 754-760. 10.2337/diacare.18.6.754.View ArticlePubMedGoogle Scholar
- Snoek FJ, Pouwer F, Welch GW, Polonsky WH: Diabetes-related emotional distress in Dutch and U.S. diabetic patients: cross-cultural validity of the problem areas in diabetes scale. Diabetes Care. 2000, 23: 1305-1309. 10.2337/diacare.23.9.1305.View ArticlePubMedGoogle Scholar
- Bradley C: Diabetes Treatment Satisfaction Questionnaire (DTSQ). Handbook of Psychology and Diabetes: A Guide to Psychological Measurement in Diabetes Research and Practice. 1994, Chur, Switzerland: Harwood Academic Publishers, 111-132.Google Scholar
- Rubin DB: Multiple imputation for nonresponse in surveys. 1987, Hoboken, NJ: WileyView ArticleGoogle Scholar
- Haukoos JS, Lewis RJ: Advanced statistics: bootstrapping confidence intervals for statistics with "difficult" distributions. Acad Emerg Med. 2005, 12: 360-365. 10.1111/j.1553-2712.2005.tb01958.x.View ArticlePubMedGoogle Scholar
- The pre-publication history for this paper can be accessed here:http://www.biomedcentral.com/1471-2296/11/35/prepub
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.