The programme has been divided into 3 phases:

Phase 1 involves a cross-sectional mapping of the frequency and pattern of unscheduled and urgent care in patients with 3 long term conditions: diabetes, rheumatologic disorders and chronic obstructive pulmonary disease (COPD) over a 12 month period. It also involves the standardization study of the screening for somatization Patient Health Questionnaire 15 (PHQ-15) for use within Greek patients with long term conditions.

Phase 2 involves in conjunction with Phase-1 the development of a ‘clinical prediction rule’ which will identify patients at risk of becoming high users of unscheduled and urgent care. It also involves the testing of the validity and utility of the clinical prediction rule.

Phase 3 involves the development and evaluation of a low intensity intervention which will be specifically designed to be used in conjunction with the clinical prediction rule.

 

In all 3 phases the qualitative work will ensure user and carer feedback to the programme as it develops. The main functions of the qualitative work will be to:

•  triangulate with the quantitative methods on the research questions,
• identify mechanisms accounting for quantitative findings,
• inform implementation of interventions, and
• inform future commissioning of our pilot interventions into practice.

 

Definition of Frequent Unscheduled or Urgent Care

AED in Greece provide services to patients seeking both Unscheduled and Urgent care.  Thus, we will refer to both terms as similar. In the early part of the programme, we will work in collaboration with the Director of the AED of the University General Teaching Hospital of Ioannina and with the GP in charge to derive a working definition of frequent unscheduled or urgent care. As yet, there is no clear and widely accepted method for defining frequent unscheduled care. In the project we may use the number of visits per year as a continuous variable and categorize patients to user groups according to median, IQR (i.e., 4 groups: (A) ≤ 25 IQR; (B) > 25 IQR – median; (C) ≥ median – 75 IQR; and (D) ≥ 75 IQR).

 

PHASE 1:Mapping out the problem

Objectives:

• To derive estimates of the frequency and pattern of unscheduled or urgent care in patients with: diabetes, rheumatologic disorders and COPD as examples of common long term conditions (Study 1A).
• To standardize the screening for somatization Patient Health Questionnaire 15 (PHQ-15) for use within Greek patients with long term conditions (Study 1B).

 

Study 1A : The Ioannina ‘clinical prediction rule’ study – Part I

We will undertake a cohort study of patients with long term conditions who seek care in the A & E Department of the Ioannina General Teaching Hospital during a period of 12 months (cross-sectional study) and we will proceed prospectively to follow-up this cohort for an additional 12-month period (Study 2A – The Ioannina ‘clinical prediction rule’, Part II). The first cross-sectional study will enable us to obtain specific and detailed information about the proportion of patients with each of the three conditions who access unscheduled or urgent care over a 12 month period and subsequent hospitalisation.

At this stage of the programme we will interview using the Greek version of the Mini International Neuropsychiatric Interview (MINI) people with each of the target conditions who visit our A&E Department in a 24-hour basis seeking urgent care. Evidence from other countries’ A & E departments showed that patients with these conditions constitute approximately a percentage of 15-20% of the total urgent visits, and it is reasonable to assume that these percentages could be higher in Greece, given the nature of the care provided by our A&E departments.

In addition, we will administer the baseline questionnaires with full explanation of the project and a consent form that requests permission to examine the patient’s medical records after one year to identify all scheduled and unscheduled visits for 1 year after the questionnaire is administered.

The baseline questionnaires will include measures of socio-demographic data, psychosocial status, and severity of illness (both generic and specific to the particular condition). The battery of the questionnaires for the assessment of psychological status will include instruments with known applicability in our research population based on our previous research in the field, including instruments assessing a wide range of psychological distress symptoms (SCL-90-R, PHQ-9 and PHQ-15), illness perceptions (B-IPQ), coping with health stressors (SOC), defensive profile (DSQ) and health-related quality of life (WHOQOL-BREF). The measures will be administered to patients over a 12 month period and then they will be followed up over the subsequent 12 months. This will establish the variables that predict frequent use of unscheduled care.

 

Study 1B: Standardization of the PHQ-15

Based on self-report questionnaires administered in the Study 1A, we will test the internal consistency and convergent and concurrent validity of the Patient Health Questionnaire 15 using the method developed by our research team for the assessment of the psychometric properties of the PHQ-9 for use within Greek patients with rheumatologic disorders.

 

PHASE 2:Identifying which people to target

Objectives

• To fully develop and test a clinical prediction rule that will identify patients with long term conditions who are at risk of becoming frequent users of unscheduled or urgent care.
• To identify personal reasons for use of unscheduled or urgent care including barriers to access for routine care, patients’ motivations, expectations and decision-making processes, and influences from families and relevant healthcare workers.

 

Study 2A: The Ioannina ‘clinical prediction rule’ study – Part II

After one year we will administer a further brief questionnaire to all participants to assess:

• health status using generic and disease-specific measures
• details of any use of unscheduled care
• reasons relevant to the patient's decision to use unscheduled care

Service use including unscheduled or urgent care will be assessed for the 12 months following entry into the study. Data will be obtained by re-approaching and re-interviewing the patients participated in the cross-sectional study either at the hospital or by telephone and from the local hospital databases concerning contacts, and hospital emergency services.

 

Statistical analysis plan to identify predictors of frequent unscheduled care.

A clinical predictor rule for frequent users will be developed using logistic regression employing a variety of selection algorithms in the search for effective explanatory variables. The performance of the models will be evaluated through estimation of sensitivity, specificity and predictive values. Internal validation of the model will also be performed using an appropriate statistical method (i.e., boot-strapping analysis). 

 

Study 2B: Why do people with chronic disease use unscheduled care?

Objectives

To identify personal reasons for use of unscheduled care including barriers to access for  routine care, patients’ motivations, expectations and decision-making processes, and influences from families, and relevant healthcare providers. Our intention is to identify processes, which might be accessible to intervention, that explain why psychosocial factors lead to patients seeking unscheduled care. This study will involve analysis of information from key informants as part of the qualitative theme of the programme.

Methods

Patient sample. Using the criteria and procedures of Studies 1A & 2A, we shall purposively sample from each disease group about 10 high users, using the definition of frequent users established by the preliminary work in study 2A. We will identify high users who have accessed unscheduled or urgent care within the previous 6-month period. We will also select 5 low users in each disease group for interview about how they manage the kinds of factors that prompt unscheduled care in high users. Recruitment will continue in parallel with analysis to allow modification of the interview guide as necessary to clarify and test the analysis.

Patient interviews. For high-users, interviews will identify reasons for their use of such care by prompting them to describe: their illness-perceptions; their motivations and expectations in accessing unscheduled care; contextual influences on their decision-making; and their experience of unscheduled care. In addition, interviews will address instances resembling those which prompted care, but where patients did not use care.

Analysis. Patients’ accounts will be analysed in parallel, using the principles of Framework Analysis. Within and between cases, particular attention will be given to contrasting instances where care was used or avoided. Content analysis will produce a categorisation of factors (at levels of patient, service, social and cultural) that promote or deter unscheduled care. We will note similarities and differences between users with different chronic conditions. One purpose of this study is to understand overarching and common themes of users with chronic illness, so that meta themes, generalisable to other illness groups can be identified, as well as specific disease related factors.

 

PHASE 3: Developing an appropriate intervention

Objectives

• To develop and evaluate an intervention which will reduce unscheduled and urgent care, and be acceptable to patients and health professionals, whilst maintaining or improving patient benefit.
• Our intention is to develop a preventative intervention which could be used in conjunction with the clinical prediction rule to reduce unscheduled and urgent care in patients with long term conditions. We intend to develop a low intensity intervention which is practice based and capable of being delivered to a large number of patients, so all people identified as potential high users could receive it.

 

Study 3. An intervention to reduce high use of unscheduled or urgent care

Development of the Intervention

In determining the nature of the intervention we will consider the following: potential therapeutic benefit; simplicity; generalisability; and sustainability. We cannot be specific about the nature of the intervention at this stage, but it is likely to involve a whole systems approach, with potential changes at the level of the consultation and practice. We are not seeking in this part of the programme to develop a new intervention but to build on and refine known evidence based psychosocial interventions such as CBT based interventions (including problem solving and cognitive strategies to enhance self management techniques), lifestyle changes including, diet and exercise. Our focus will be to focus on the acceptable, accessible and feasible components of these interventions and maximise patient choice and preference.

Low intensity self management strategies are likely to be offered in a range of formats (including for example a book, CDROM, or via a website depending on patient preference). Other key components of the intervention including level of support by a health professional will be determined in the development phase.

 

Stated Preference Methods

We will use stated preference methods (e.g. discrete choice experiments) to explore preferences for different process and outcome attributes of treatment options. This will enable us to investigate the acceptability of particular treatment options to users. Using the stated preference survey, participants will be given a series of hypothetical scenarios and asked to make choices about their preferred treatment options. The attributes and levels for the scenarios will be developed from preliminary qualitative research findings. The attributes will describe different components of potential interventions (e.g. waiting time, care setting, type of health care professional seen, type of intervention, and outcome of care).

The outcomes of this project will enable us to determine which aspects of treatment/service delivery are most highly valued by different groups of patients. We are particularly interested in tailoring interventions to suit particular user/patient characteristics. For example, psychological therapiesmaybe more acceptable to women than men.

 

Focus Groups

The data from the stated preference survey will then be combined with further data emerging from the Phase 2 qualitative studies  to  finalise the  intervention.  We recognise that the Phase 2 work will not have been fully completed at this stage, but it is likely that clear themes will have emerged that will have potential therapeutic value.

 

Intervention

• Hypothesis: There will be a reduction in unscheduled or urgent care, over a 12 month period, in patients with long term conditions in which the clinical prediction rule intervention is implemented in comparison to those without the intervention.
• Methods: We will seek approval from our local medical ethics committee to interview/assess  patients to assess the feasibility of implementing the clinical prediction rule intervention, and potential benefits from it. We will introduce the intervention over a six month period. It will be supported by web-based resources and computer templates/prompts.  We will involve users in the training process. We will also ask patients to complete psychosocial assessments and measures of health status on a sequential basis every 3 months once they have entered the pathway, to gain an indication of health benefit.
• Statistical considerations: The main indicator of the success of the intervention will be the total number of unscheduled visits by patients identified as having long term conditions recorded in the follow-up period.

 

Workpackages

Work Package 1

• WP1a.  Estimation of the frequency and pattern of unscheduled and urgent care in Greek patients with diabetes, rheumatologic disorders and chronic obstructive pulmonary disease (COPD).
• WP1b.  Identification of the psychosocial factors associated with unscheduled and urgent care in patients with these 3 long term conditions.
• WP1c. A standardized version of the screening for somatization instrument Patient Health Questionnaire 15.

 

Work Package 2

• WP2a. A clinical prediction rule for identifying patients who are potential high users of unscheduled care.
• WP2b. Identification of key elements that can be influenced by new interventions to reduce urgent care.
• WP2c. A rich detailed and highly practical understanding of who accesses unscheduled care, why they do this, what they experience, and what happens to them.

 

Work Package 3

• WP3. Our programme is expected to demonstrate the feasibility of introducing the clinical prediction rule care intervention into AED. Other tangible outputs will include a robust system for detecting potential high users of unscheduled care in long term conditions and a user informed intervention to be used in conjunction with the clinical prediction rule for the prevention of unscheduled care. The expectation of the research team is that work from this study will inform clinical practice and policy.