• Introduction
• Preparations
• Protocol
  • Abstract
  • Background
  • Objectives
  • Methods
  • Analysis
  • Ethics
  • Organisations and support
  • References
  • Research questions
• Training
• Procedure
• A Patient's view

• 
• Print

Analysis

The data will be analyzed using SPSS. Descriptive statistics, correlations, analysis of variance (ANOVA) and multiple regression analysis techniques will be applied to answer the Research Questions.

The topics/themes as indicated by patients will be summarized using descriptive statistics and correlated with patient demographics (age, sex, ethnicity), clinical parameters (duration, type diabetes, HbA1c, complication status) and well-being outcomes (PAID, WHO-5).

To determine the level of well-being in the group as a whole, mean WHO-5 and PAID scores will be calculated on Time 1 and Time 2, and in sub-groups: males and females, type 1 and type 2 diabetes patients, with 0 vs. 1 or 2+ complications, with/without co-morbidity, with/out history of mental health care. Mean differences between T1 and T2 on WHO-5 and PAID will be tested using analysis of variance, using p<0.05 as level of significance.

The percentage of patients reporting poor/moderate/good psychological state on Time 1 based on the WHO-5 scores (= 28, > 29 < 50, and = 50) and percentage patients reporting low/average/high levels of diabetes-related emotional distress, based on PAID scores (= 20, > 20 < 40, = 41) is calculated. This will allow for identifying ‘caseness’ and specific problems areas (items, subscales). The characteristics of patients in poor vs. good well-being are analyzed to test for socio-demographic, life events and medical-biographic differences/profiles, both on T1 and T2.

The percentages of patients reporting to currently receiving mental health care are determined for the group as a whole and for the subgroup of patients reporting low well-being (WHO-5< 50 and < 28 respectively, and PAID> 40) on T1 and T2. This information provides insight into the degree to which psychological needs of patients are met and whether changes have occurred in between T1 and T2.

At T1 and T2 the frequency of areas of concern to the patients, as indicated on the assessment form (1 item) is calculated and cross-tabulated with poor vs. moderate to good scores on well-being indices. This will provide insight in the concordance between patients’ indication that mood/stress is an item to be discussed and questionnaire scores.

Pearson correlations will be calculated to test for significant associations (P <0.05) between WHO-5 and PAID scores on T1 and T2 on the one hand and demographic (gender, age, education, living alone/together, work) and medical characteristics on the other. Of particular interest is the association between psychological well-being and glycaemic control.

Using a cut-off of HbA1c > 8% for poor control and WHO-5 score <28 for poor well-being, a 2x2 matrix can be calculated showing the percentage of patients in good well-being/poor control, good well-being/good control, poor well-being/poor control, poor well-being/good control. Similar analyses are performed using PAID score > 40 as marker for high diabetes-specific distress. Demographic and clinical characteristics of these sub-groups are explored. Changes between T1 and T2 are documented.

PRISM data will be analysed secondary, mainly for psychometric purposes.

Process data will be analysed on 3 levels: patient satisfaction, registration of actions directly following the assessment of well-being, and the clinical experience.

Level 1: Patient satisfaction will be investigated after the 2nd assessment, i.e. after 12 months, using a two-page, anonymous questionnaire, asking patients about their experiences of the monitoring procedure. This evaluation questionnaire is handed out to the patient with a pre-stamped envelope, to be returned within two weeks.

Level 2: The actions taken directly following the assessment procedure will be registered by indicating the number of referrals to mental health services, either within the hospital or to external services. Also, the number of extra consultations with a health care team member (doctor, nurse, dietician etcetera) as a direct consequence of discussing well-being will be recorded.

Level 3: The participating clinical team members will be invited to fill out an evaluation form pertaining to their clinical experience (e.g. discussing outcomes, impact on time). In addition, interviews with a sample of centres will be held to specify the benefits and barriers of the monitoring procedure. The coordination of the evaluation will be at the VUMC.

The questionnaires on patient satisfaction and clinical experience will be developed in due course.

Back to top