Baseline data of a population-based cohort of patients with diabetes in Switzerland ( CoDiab-VD )

QUESTIONS UNDER STUDY: To describe a populationbased sample of patients with diabetes and the quality of their care in the canton of Vaud, Switzerland, as a baseline measure for the evaluation of the “Programme cantonal Diabète”. METHODS: We conducted a self-administered paperbased questionnaire survey. Noninstitutionalised adult (aged ≥18 years) patients with diabetes diagnosed for at least 1 year and residing in the canton of Vaud were recruited by community pharmacies. Women with gestational diabetes, people with obvious cognitive impairment or people not sufficiently fluent in French were excluded. Primary outcomes were recommended processes-of-care and outcomes of care (glycosylated haemoglobin [HbA1c], generic and disease-specific health-related quality of life (HRQoL), overall care score in relation to the Chronic Care Model). Other measures included diabetes education, selfmanagement support and self-efficacy, health status, health behaviour and demographics. RESULTS: A total of 519 patients with diabetes were included. Whereas the mean HbA1c level was 7.3% (n = 177, 95% confidence interval 7.1–7.5), diabetes-specific processes-of-care and influenza vaccination were reported by less than two-thirds of the patients. Physical activity and diet recommendations results mirrored patients’ difficulties with their management in daily life and diabetes-specific HRQoL was worst in the dimensions relative to diet (eating and drinking) and sex life. A minority of patients reported ever having participated in diabetes education courses (32.8%). Overall, patients were satisfied with their care and the support they received. CONCLUSIONS: This study provides a broad picture of the experiences of people living with diabetes in the canton of Vaud. It shall guide the development of targeted interventions within the “Programme cantonal Diabète”. Trial registration number: ClinicalTrials.gov, identifier NCT01902043.


Introduction
Worldwide more than 370 million people had diabetes in 2012 [1].Its occurrence continues to increase, with debilitating clinical consequences [2,3].Despite a universal healthcare system, Switzerland is not protected against this upward trend.The overall prevalence of diabetes in Switzerland was reported to be 6.1% in 2010 [4] and 6.6% in 2006 in the canton of Vaud, Switzerland [5].However, only scarce data are available regarding patient characteristics and the quality of their care [6].Indeed, previous Swiss data showed that there was room for improvement regarding diabetes care and adherence to diabetes guidelines [7][8][9].Although diabetes-specific processes-ofcare were not so often performed by primary care physicians (two-thirds of patients had had annual eye and foot screening, 50% had had annual renal assessment) [8], this was somewhat better than initially measured [7].Population-based studies are essential when studying chronic diseases, as they can offer an overall understanding about the trends in the prevalence of diabetes and the state of diabetes care in specific settings [10].These studies help providers and public health advocates better to explain the current challenges of communities living with diabetes, and to tailor successfully supportive interventions [11].In the past, population-based studies on diabetes had primarily focused on sociodemographic factors, medical outcomes, health services and quality-of-care indicators [12][13][14].Now, there is a growing interest in using patient-reported outcome measures [15], with emphasis on patient-reported health status, health-related quality of life (HRQoL) and self-management [1,16].These latter measures are increasingly important as we now know that patients actively involved in their care are more engaged, better informed and benefit from improved health outcomes [17].Indeed, patients who express higher HRQoL function better within the healthcare system and perform more health-promoting activities [18,19].Towards this goal, patient-reported outcomes, in combination with other quality-of-care indicators, are necessary for healthcare systems to better understand and meet the needs of patients [10].
In 2010, based on the Chronic Care Model, the Public Health department of the canton of Vaud created a pioneer programme in Switzerland, the "Programme cantonal Diabète", to reduce the impact of diabetes on the population, specifically by (1.) acting on prevention, and (2.) improving care for people living with diabetes [20].In this paper, we aimed to present a comprehensive picture of the health and care of a population-based sample of patients with diabetes.More specifically, we wanted to consider not only the self-report of usually considered primary processes and outcomes of care (e.g., HbA 1c , HRQoL), but also of other daily life outcomes of importance to the patients such as health status, health behaviours, diabetes education, selfmanagement support and self-efficacy.These data will represent the baseline measures of CoDiab-VD, a populationbased cohort study on diabetes care in the canton of Vaud, Switzerland.

Study population
We conducted a paper-based survey with recruitment in two waves (autumn of 2011 and summer of 2012).Patients with diabetes, consecutively recruited through communitybased pharmacies in the canton of Vaud, were invited to participate if they visited an included pharmacy with a prescription for diabetes-related treatment and/or equipment (oral antidiabetic medications, insulin, glycaemic strips or glucose meter).Eligible participants were noninstitutionalised adult (aged ≥18 years) patients with a diabetes diagnosis for at least 1 year and residing in that canton.Women with gestational diabetes, people with obvious cognitive impairment, or people not sufficiently fluent in French were excluded from the survey.Pharmacists checked patients' eligibility and then briefly presented the study; patients who agreed to receive the questionnaire package (information letter, questionnaire, prepaid reply envelope) were asked to complete it at home and to send it back to the investigators.The total number of eligible participants and the characteristics of nonparticipating patients were not recorded by pharmacists because of time and organisation constraints.The recruitment was conducted without follow-up recall [21].Sample size was based on estimates of good precision (i.e., confidence interval width) around the following primary outcomes: mean glycosylated haemoglobin (HbA 1c ), mean Physical and Mental component scores (PCS and MCS) of the Short Form-12 Health Survey (SF-12), Patient Assessment of Chronic Illness Care (PACIC) global score, and percentage of patients receiving recommended annual processes-of-care.Given pharmacy clusters (40 pharmacies, each recruiting 15 patients, intraclass correlation 0.05, alpha 0.05, beta 0.2), 600 participants were targeted.The protocol of this study was approved by the Cantonal Ethics Committee of Research on Human Beings of the Canton of Vaud (Protocol N°151/11).Written informed consent was obtained from all participants, and data were kept anonymous and confidential.

Data collection
Participants completed a self-administered paper questionnaire.Primary outcomes of interest were recommended processes and outcomes of care indicators.The processesof-care indicators considered were: HbA 1c check among patients who were aware of what HbA 1c is ("HbA 1caware"), blood pressure measurement, weight measurement, lipid profile, annual diabetic foot examination, annual urine test for microalbuminuria, annual eye examination by ophthalmologist, seasonal influenza vaccination, physical activity and diet recommendations.Outcomes of care were: mean HbA 1c among HbA 1c -aware patients, generic and diabetes-specific HRQoL measures (SF-12 PCS and MCS, score range 0 = worst to 100 = best; Audit of Diabetes-Dependent Quality of Life 19 -ADDQoL, range -9 = maximum negative impact of diabetes to +3 = maximum positive impact of diabetes), and congruency of care with the Chronic Care Model (PACIC score, range 1 = lowest to 5 = highest congruency).We also considered the following diabetes education, self-management support and self-efficacy measures: glucose self-monitoring, HbA 1c knowledge, participation in diabetes education course, membership in the local diabetes association (Association Vaudoise du Diabète -AVD), knowledge about the "Diabetes Passport" (a small booklet, kept by the patient, in which consultations and medical results are noted), easiness-difficulty with daily self-management of diabetes in general, with physical activity, medication and diet, as well as their level of information about diabetes and source of information, support and satisfaction with healthcare team and personal social network.Other data included age, gender, sociodemographics (relationship status, living situation, family income in Swiss franc (CHF/month), education, employment status, type of residence, nationality), and diabetes characteristics (type, duration, treatment, complications), health behaviours (smoking status, alcohol consumption, physical activity), and comorbidities.

Data analysis
Descriptive analyses were conducted, with data reported as means or percentages for continuous or categorical variables, respectively.For primary outcomes, confidence intervals were also calculated, taking into account the hierarchical structure of the data (clusters of pharmacies).All analyses were performed using Stata 12.1.

Patients' characteristics
Out of 1,054 eligible people who accepted a questionnaire from pharmacists (809 in 2011, 245 in 2012), 519 patients completed and returned the questionnaire (406 in 2011, 113 in 2012, participation rate of 50.2% and 46.1%, respectively).Table 1 shows participants' characteristics.The majority were male (59.7%) and mean age was 64.5 years.Most reported a diagnosis of type 2 diabetes (66.9%), and 48% of respondents had been diagnosed over 10 years ago.Although less than 20% of patients reported being current smokers, almost 50% presented high-risk drinking beha-

Original article
Swiss Med Wkly.2014;144:w13951 Swiss Medical Weekly • PDF of the online version • www.smw.chviour, more than 80% were either obese or overweight, and 30% were considered to be inactive physically.

Primary outcomes
Blood pressure, weight, and lipid profile were measured in over 90% of participants in the prior 12 months (table 2).HbA 1c was checked at least once in the last year in 98.3% of HbA 1c -aware patients (n = 282).Other processesof-care indicators were reported by less than 70% of the patients.Low percentages were also found for physical activity and diet recommendations.Among HbA 1c -aware patients, mean patients-reported HbA 1c was 7.3% (n = 177).Whereas generic health-related quality of life results showed worse physical than mental health (SF-12 PCS: 43.4,SF-12 MCS: 47.0), disease-specific HRQoL showed that patients reported being most affected by limitations in their daily life regarding "freedom to eat", "sex life", and "freedom to drink", while "people's reaction" and "financial situation" mattered least (fig.1).Mean patient assessment of chronic illness care (PACIC) score was 2.8 (1 = lowest to 5 = highest congruency).

Diabetes education, self-management support and selfefficacy
Home glucose self-monitoring was frequently used (81.6%)(fig.2A), but HbA 1c knowledge (58.3%) and participation in a diabetes education course (32.8%) were rather low.Only 14.3% were members of the local diabetes association and 18.4% knew about the "Diabetes Passport".Measures of "self-efficacy" showed that diet posed a great- er difficulty when compared with daily general diabetes management (fig.2B).In addition, the majority of patients felt well-to very well-informed about their chronic disease (85.9%), with the major provider of information being their medical doctors (92%).Other sources of information were media (30.8%), other healthcare professionals (23.5%), social network (8.6%), association or healthcare network (5.7%), alternate sources (3.9%) and none (1.8%).Finally, 67% of the participants were well-or very-well satisfied about their care, and nearly 58% would definitely recommend the care received (fig.2C).Most were satisfied by the support obtained from their healthcare team (72.6%) and members of their social network (72.3%).

Discussion
In this survey, we comprehensively described the care of a cohort of 519 patients with diabetes in the canton of Vaud, Switzerland.Our results signalled that quality gaps remained, especially for certain diabetes-specific quality indicators.We also noted a disconnection between patients' perceptions and their actual disease knowledge and care; indeed, patients felt satisfyingly informed about diabetes, but yet many reported not even knowing what HbA 1c meant.Although satisfaction with and support from healthcare providers were rated high and the majority of participants would recommend their healthcare, recommendations about physical activity and diet were well below goal.This is doubly significant as patients expressed their greatest management difficulties in these areas.
The percentage of patients reporting uncontrolled diabetes (5.7%), defined by a HbA 1c over 9%, as well as the percentage of patients with HbA 1c level below 8% (73%), were below and above, respectively, the 15% and 60% cutoff recommended by the Swiss society of endocrinology and diabetology (E.Christ, personal communication).The mean self-reported HbA 1c value was good (7.3%), but only available for the 177 HbA 1c -aware patients.Mean HbA 1c levels for the whole sample would probably be less satisfactory.Generic HRQoL results showed that patients reported better MCS than PCS scores; this is similar to previous population-based studies on diabetes using the SF-12 scoring [22,23].Diabetes-specific HRQoL measures (ADDQoL) showed that "freedom to eat" was a quality-oflife domain where patients felt particularly affected.This difficulty, also reported by Collins et al. [24], echoed the fact that patients reported more difficulties when it came to the management not only of diet but also of physical activity.These findings point to the need for more emphasis on guiding patients on these topics since patients reported few recommendations regarding both diet and physical activity.Increased participation in diabetes education programmes and diabetes association could be a targeted way to address these needs since higher levels of self-management and self-efficacy have been shown to have a positive impact on diabetes-related outcomes such as HbA 1c [25].
Another way to improve patient' knowledge, attitudes and behaviour about nutrition could be to increase physicians' referral to dieticians and specialised nurses.This would be particularly appropriate since our participants reported un-

Original article
Swiss Med Wkly.2014;144:w13951 social network or healthcare providers possess more selfmanagement skills and self-efficacy regarding their diabetes, and feel more confident about their ability to complete health-promoting activities [27,28].Planning activities for patients with diabetes is important.All healthcare providers should be involved in promoting these resources to patients.The collection of a broad range of outcomes and measures is an added value of this study.However, some limitations must be taken into consideration.First, these results are based on patients' self-report which may be prone to recall bias.However, supplementary analyses demonstrated good correlation between patient-reported data and physician records for simple processes-of-care, when data was supplied by the treating physician for a fraction of the cohort (H.Collet, personal communication); in addition, these results were similar to those of other population-based studies [29].Also, despite the possibility of under-or over-estimation of self-report indicators of care, patient-reported outcomes are important measures to consider when evaluating the health status of a population and its needs regarding healthcare [15].Second, 519 patients were recruited in the study instead of the 600 targeted.As we had more clusters (pharmacies) than expected, and because of our conservative sample size calculations, the precision around point estimates was nevertheless acceptable.Finally, our results may not be generalisable to the whole population of patients with diabetes in the canton of Vaud, as only patients fluent enough in French to fill in a questionnaire, and with a diagnosis of diabetes for over a year and visiting a pharmacy, were recruited.We cannot exclude that results might have been different with the inclusion of patients speaking foreign languages such as migrants or highly-skilled and educated English-speaking expatriate workers.However, the direction and size of the effect of a broader inclusion of patients may not be so straightforward to apprehend.Also, participants showed similar characteristics in terms of age, gender, education and BMI than those from prior studies in the same canton [6,30].In addition, they represent a population-based sample of patients with diabetes more closely than participants recruited in hospitals [9].These findings will be useful for the further development and evaluation of the "Programme cantonal Diabète".They are particularly interesting since they constitute the baseline data of the cohort of patients with diabetes in the canton of Vaud (CoDiab-VD).On a more global level, these results shall support institutional strategies to close knowledge gaps for patients with diabetes.More specifically, they should ensure that all processes-of-care are received by patients -especially those specific to diabetes care, and that diabetes education programmes are promoted and made accessible to all patients.A special emphasis should be given to diet and physical activity education and support.As this study covered a wide spectrum of quality of care indicators, these results should be of interest to healthcare stakeholders seeking to improve care for patients with diabetes both in Switzerland and elsewhere.

Figure 1 Figure 2
Figure 1 Average weighted impact score of the 19 domains of the ADDQoL.ADDQoL = Audit of Diabetes-Dependent Quality of Life

Figure 1
Figure 1 Average weighted impact score of the 19 domains of the ADDQoL.ADDQoL = Audit of Diabetes-Dependent Quality of Life

Figure 2 Self
Figure 2 Self-management support.A. Knowledge and information.B. "Self-efficacy" measures.C. Social support and satisfaction about it.AVD = Association Vaudoise du Diabète; HbA 1c = glycosylated haemoglobin

Table 1 :
Baseline characteristics of the participants (n = 519 patients with diabetes).