Purpose  
The aim of the proposed study is both to identify and quantify the relationships between patient outcomes and the perceived obstacles to adherence of people living with type 2 diabetes who receive diabetes care in primary care settings.

Background of the project
The researchers have already completed a comprehensive search of the literature on adherence (1) and a qualitative study in 7 European countries recently completed by the applicants (2,3,4). These revealed the obstacles to adherence to treatment recommendations for patients living with type 2 diabetes. There is clearly a link between adherence and successful diabetes management.  Although we now know what the obstacles are, we do not know which obstacles have the biggest effect on outcomes. Thus we lack evidence to guide us on how to prioritise which obstacles to address in order to have most benefit for the patients.  We do not know whether patients who experience either of these obstacles actually have worse outcomes than patients who do not. So, we do not know whether addressing those obstacles is useful. 

As no appropriate, comprehensive instrument exists, we first need to validate the instrument (the OBO questionnaire). The source of the questionnaire items are the findings of focus group discussions on adherence to treatment recommendations in seven European countries (EUROBSTACLE study).

We acknowledge that many instruments are published, but none covers all the obstacles identified by patients in our previous studies, nor provides enough detail. We then aim to link the obstacles experienced to diabetes outcomes, to provide an evidence based guide on the importance of each obstacle. We shall collect outcome data from patients’ GP records on physiological measures and from a quality of life questionnaire. The relationships between these measures and the experience of obstacles, of the same patients, will be systematically analysed. This will enable the researchers to prioritise future attempts to reduce obstacles (by research and service improvement) based on which obstacles relate most to poor outcomes and by how much. 

Our previous work (3) was partly supported by the European General Practice Research Network. We found that the obstacles to adherence were common across all 7 countries despite the variations in health care systems. Hilary Hearnshaw and Kay Wright have created, from these results, the OBO questionnaire to measure the obstacles experiences by patients.    

Plan of investigation

In Phase I the OBO questionnaire will be validated. In Phase II the main survey will be conducted using the OBO questionnaire. Phase I will also be a feasibility pilot for Phase 2 study, in which data will be gathered on recruitment, questionnaire response rates, the feasibility of collecting physiological data, and the feasibility and validity of the proposed methods for analysis and interpretation of the data. This will also inform the sample size required and the conduct of Phase 2.

 
Methods
1. Phase 1: Questionnaire validation

Questionnaire development

The obstacles to adherence found in our previous study (2,3) can be divided into five categories: diagnosis, knowledge, treatment, communication and motivation. The items within these themes were the components of the OBO questionnaire. We used the literature on questionnaires related to diabetes to select appropriate words and phrases relevant to themes identified in the study with focus groups. We also checked for items which were in the literature, but not on our item list, and added them in.
The OBO questionnaire is designed to measure obstacles experienced by the respondent in living with type 2 diabetes. The questionnaire has been amended and approved by the Warwick Diabetes Care user Group (5). The criterion validity of the OBO questionnaire will be assessed by comparison with the gold standard in this field, which is the Problem Areas in Diabetes (PAID) questionnaire (6). The PAID question will be included in the study questionnaire for this purpose. Permission from its authors has been granted. The ADDQoL (perceived impact of diabetes on quality of life) (7) will be used to measure quality of life, specific to each respondent.

Ethics committee approval should be obtained.

The OBO questionnaire has been written in English. It will be translated in Dutch and Estonian.  To check the quality of the translation, the questionnaire will be translated and retranslated by two different people.

 
Sample
We estimate the necessary sample size for validity testing to be 300 patients (8): 100 on each of 3 countries. Patients will be recruited by 20 GPs or nurses (in the UK), each inviting consecutive patients, who have a confirmed diagnoses of type 2 diabetes, to participate until 7 have agreed. Patients will also need to be able to complete a questionnaire in the native language (English, Dutch or Estonian). Anticipating a refusal rate and a drop out rate of 30% this means that each GP or nurse will probably have to invite 10 people in order to get 7 to agree and 5 to actually return the questionnaire. We anticipate that GPs and patients can be recruited within a period of 2 months. Data collection should be completed within 3 weeks of the last recruited patient. 

Inclusion criteria for GPs: Each country decides on strategies. No representative sample is required, GPs may be recruited from the research group in contact with the research centre.

Inclusion criteria for patients: consecutive patients with established type 2 diabetes

• study inclusion not offered
• patient refusal: - practical reason (language, reading,…)
• does not want to participate
• not interested
• no time
• not at least one eligible outcome measure in last year’s record 

Patient characteristics:

• age
• gender
• duration since diagnosis: exact duration or unknown

Consent forms:

• participation in study
• extraction of data from doctor’s record

Instruction letter for GP:

• type 2 diabetes patients
• consecutive patients
• for every, consecutive, patient with type 2 diabetes fill in an inclusion document 

Measures
We shall test the feasibility of collecting the required data. The measures of patient outcomes will be physiological (HbA1c, total cholesterol/HDL ratio, blood pressure, BMI) and the quality of life (ADDQoL) (7). The analysis of the data from these measures will explore any relationship between the OBO questionnaire scores (dependent variable) and the physiological and quality of life measures. Multilevel regression analysis will be used.

  
At the practice after signing consent, participants will be given the OBO questionnaire to complete. The questionnaire will be returned tot the researchers using a prepaid envelope. The GP will provide physiological outcomes from their records (at least one HbA1c, the most recent serum cholesterol and serum HDL value, blood pressure and Body Mass Index or height and weight data) and note them on a checklist.

This will entail 30 visits to practice for recruitment and data collection, and 200 questionnaires given out.

Questionnaires:
ADDQoL should be validated in Belgium and Estonia. The PAID questionnaire has already been validated in Dutch. Whether this questionnaire has been validated in Estonian is not known yet.

An option is to ask 50% of the patients to fill in the PAID and 50% the ADDQoL.   

Analysis
In Phase 1, the OBO questionnaire will be tested, amended and validated. Distributions of scores will be used to identify redundant and meaningless items of the questionnaire. Factor analysis will be used to identify the valuable factors of the questionnaire, and which summary scores over groups of items can be used in the regression analysis.

 
2. Phase 2: Main OBO study
Study in many European countries with the OBO-2 questionnaire, with a sample size allowing for correlation testing. Cfr draft version 11/2003 (on file).

Timetable

Target Research Period: 10 months, September 2004-August 2005.

Month	Activity
0-3	Recruitment strategy Development of measurement instruments
3	ethics approval appointment of junior researchers
4-6	recruitment of GPs recruitment of patients
5-8	Data collection Data analysis
8-10	Writing up this study Writing next proposal

A meeting in Warwick of a researcher from each country is planned in May 2005.

Outcomes of the study

References

1. Vermeire E, Hearnshaw H.M., Van Royen P, Denekens J, Patient adherence to treatment: three decades of research. A comprehensive review. J Clinical Pharmacy and Therapeutics, 2001:331-342.
   
2. Vermeire E, Van Royen P, Coenen S, Wens J, Denekens J. The adherence of type 2 diabetes patients to their therapeutic regimens: a qualitative study from the patient’s perspective. Pract Diab Int 2003;20(6):209-214
   
3. Hearnshaw, H.M., Vermeire E, Wens J, Levasseur G, “Living with Type 2 Diabetes can be difficult in several European Health Care Systems”.6^th International Conference of Primary Care Diabetes Europe, Paris, France, 23-24 August 2003.
   
4. Hearnshaw H.M., Vermeire E, A taxonomy of diabetes care services in European countries. European Region, World Organization of Academies of General Practice Annual Congress, London, 9-13 June 2002.
   http://www2.wawick.ac.uk/fac/med/diabetes/research/user
   
5. Welch G, Weinger K, Anderson B, Polonsky WH. Responsiveness of the Problem Areas in Diabetes (PAID) Questionnaire. Diabetic Medicine 2003; 20:69-72
   
6. Bradley C, Todd C, Gorton T, Symonds E, Martin A, Plowright R. The development of an individualised questionnaire measure of perceived impact of diabetes on quality of life: the ADDQoL, Quality of Life Research, 1999; 8:79-91.
   
7. Kline P. A handbook of test construction. London: Methuen 1986.
   
8. Machin D, Campbell MJ, Payers P, Pinol A. Statistical tables for the design of clinical studies. Oxford: Blackwell Scientific Publications, 1996.
   
9. Streiner DL, Norman GR. Health measurement scales: a practical guide to their development and use. Oxford University Press, 2004
   http://www2.warwick.ac.uk/fac/med/diabetess/research/framework/