Data Collection for Diabetes Patient Management Practices and Costs of Complications: UK Results

Supplemental information

Data collection for diabetes patient management practices and costs of complications: UK results

Copyright © 2018 Fourth Hurdle Consulting Ltd1

1. Study rationale

IMS Health has a proprietary economic model for type 1 and type 2 diabetes mellitus, the IMS CORE Diabetes Model (CDM). The model has been validated and widely published, and there is a growing literature based on diabetes care interventions evaluated with the model across the world.

The CDM model has been designed to allow calculations of the long-term clinical and economic outcomes of type 1 diabetes mellitus(T1DM) and type 2 diabetes mellitus (T2DM). It incorporates patients’ cohort baseline characteristics, past history of complications, diabetes management, concomitant medication, screening strategies, and changes in physiological parameters over time[1].

The model is based on a series of sub-models which simulate important complications of diabetes (cardiovascular disease, eye disease, hypoglycaemia, nephropathy, neuropathy, foot ulcer, amputation, stroke, ketoacidosis, lactic acidosis and mortality). Each sub-model is a Markov model using Monte Carlo simulations and incorporating time, state, time-in state, and diabetes type-dependent probabilities derived from published sources[1].

Flexible adaptations of the model to different economic settings, cohort characteristics, treatment pathways and clinical outcomes can be easily performed. This can be done through the model user-friendly interface which allows editing of the economic and clinical data. Individual country analysis can be thus performed to reflect specific economic and patients’ management settings.

The scope of the present reviewis to conduct an update of the CDM economic, patients’ management data and cohort characteristics related to smoking and alcohol consumption. The update will reflect current care patterns and direct medical expenditure associated with diabetes as per national guidelines, reimbursement levels and national or regional tariffs for each country of interest.

The update of the costs, management practices and patients’ cohort characteristics related to smoking and alcohol consumption is based on version 7.0 of the CDM. Thus, the definitions of the costs and management parameters used for the update are in line with this version of the CDM (version 7.0).The results for the UK are presented.

1. Objectives

The objectives of the presentupdate are:

• To review and collect up-to-date data on the costs of diabetes-related complications for the year 2010[1] for the UK for use in the CDM.
• To review and collect up-to-date data on patients’ management practices for the market of interest to be used in the CDM.
• To review and collect up-to date data on patients’ cohort characteristics related to smoking and alcohol consumption to be used in the CDM.

1. Methodology

In order to update the current economic, patients’ management and cohort characteristics related to smoking and cigarettes consumption, a desk based research was initiated by IMS Health on all parameters agreed to be included in the present update. The desk based research was followed by primary research through contacts with key opinion leaders (KOLs)where substantial data gaps were identified.

3.1Update of the economic inputs of the CDM

3.1.1Target population

The data collection was aimed to individualise costs derived from diabetes-specific sources where possible and in line with published evidence that the cost of treating complications in patients with diabetes may well be higher than in the general population.

Whereeconomic studies specifically conducted on diabetic patientswere not available, costs estimates on non-diabetic patients suffering from the respective disease were used.

Costs in both T1DM and T2DM have been used without distinction between the two.Where costs or patients’ management data were available for each, T1DM and T2DM, these were reported separately.

3.1.2Costing perspective

All costs considered in the present update reflect the third party payer’s perspective as either the national or regional public healthcare payer or a third party insurer.

3.1.3Definition of costs

Direct medical costs attributable to diabetes complications formed the subject of the update. Indirect costs associated with diabetes complications were not been included, and direct non-medical costs such as home care were excluded where possible.

Costs of diabetes-related complications are associated, as defined in the CDM, with either acute events (event costs) or chronic conditions of an ongoing disease (state costs). Costs related to the ongoingdisease reflect annual medical care consumption.

For chronic complications, a distinction was made betweencostsarising in the first year after disease onset and subsequent years. It is expected that some complications would be characterised by higher costs in the first year as a consequence ofhigh initial hospitalisation costs incurred during the acute phase. Only one acute event occurring during the first year is considered in the model.Follow-up costs are accounted for every year until the resolution of the particular complication.

The comprehensive list and definition of each of the model-cost input which formed the subject of the present update is presented in Table 1 below. Broadly, costs associated with preventive interventions of diabetes complications (management costs), costs of cardiovascular complications, costs of renal complications, costs of acute events, costs associated with diabetic retinopathy, costs associated with diabetic neuropathy, foot ulcer and amputation costs have been collected and updated.

Table 1: List of model economic input parameters

3.1.4Data sources

High impact sources such as: government-published tariffs, government publications, registries publications, physicians’ consortium publications or health-economic technology appraisals have been used with priority.

In addition to this, structured literature searchesonthe EMBASE and MEDLINE databases were conducted. The literature searcheshad the scope of identifying new publications or updates of the papers originally used in the model.

A global search onMEDLINE and EMBASE was initially performed in December 2010. The search waslimited to the English literature published over the period 2000 - 2010. The complete search strategy and search strings are reported in Appendix 1.The global literature search was supplemented by unstructured searches in local-language publications.

Hospital DRG tariffs and ambulatory tariffswere used to update procedures costs, event costs and costs of contacts with healthcare professionals. National pharmaceutical formularies were accessed to update the costs associated to pharmacotherapy.

3.1.5Selection of sources

As it was not the aim of the update to perform a literature review on the costs of diabetes complications, no inclusion/exclusion criteria was defined a priori. Yet, the decision when choosing one reference over another followed a structured list of considerations as follows:

• High impact sources such as official publicationsand public tariffs were used whenever available;
• It was aimed to use only studies conducted onpatients with diabetes. Where such studies were not available, studies on the general population suffering from the respective disease were used instead;
• Only studies reflecting the country specific care patterns and direct medical costs were considered;
• The most up to date and robust estimates were considered; no publications older than 2000 (date of publication) were considered;
• Published cost estimates were chosen to best reflect the definition of the CDM inputs;
• It was aimed to only include studies where the perspective was clearly defined and direct costs could be extracted;
• General approach to missing economic evidence

After the completion of the desk research, a series of data gaps were identified.

Where the literature searches did not provide enough information to cover all the necessary cost inputs for the model, the following options were considered on an individual basis and with advice from experienced IMS local affiliates:

• if a country similar enough in the organisation of care provision, healthcare expenditure and patients profiles was identified, cost data from this country was used to fill-in the missing evidence;
• where no such country was identified, a weighted average regional cost was derived and used (e.g. European average cost) or multi-country cost studies were used;
• Where no similar healthcare setting was identified and the number of missing cost estimates was high, further contacts with clinical experts to identify the missing costs were sought.

A series of assumptions based on national clinical and treatment guidelines were also required in order to be able to cover costs which were not directlyavailablefrom the literature.

3.1.7Quality control and impact of the current update

Following the completion of the data collection, a two stage quality control was performed on the country data inputs. The first check was aimed to identify any errors in the values imputed or any inconsistencies with the referred sources. A minimum of 20% of the final reported values were checked against references for the UK.

Next, for all costs included in the present update, final values were pulled together andoutliers were identified and checked against references and against the CDM definition of the corresponding health states.

3.2Update of the patients’ management input parameters and country specific cohort characteristics related to smoking and alcohol consumption

The update of the patients’ management parameters and cohort characteristics followed the same approach with respect to the patient population, identification and selection of studies and quality assurance as in the case of the economic inputs.

The comprehensive list and the definition of each of these model inputs are reported in Table 2 below.

CDM inputs related to the reduction in risk of foot ulceration with prevention programmes, improvement in ulcer healing rate if treated with novel therapies and reduction in amputation rate with foot care were kept unchanged from the initial values used in the model. These inputs were considered to be less sensitive to the provision and delivery of care, but rather depending on the specific intervention or programme.

In the same way, inputs related to sensitivity and specificity of eye screening, sensitivity of gross proteinuria (GPR) screening and sensitivity and specificity of micoralbuminuria (MAU) screening were considered dependent on the properties of the tests used (for example the Micral test for MAU screening) and less on the patient population or the different care settings. Peer-reviewed articles reporting on each of these inputs were used.

Table 2: List of model parameters on disease management and cohort characteristics related to smoking and alcohol consumption

1. Results
2. UK
3. Discounting

For the reference case an annual discount rate of 3.5% should be used for both costs and benefits. When results are potentially sensitive to the discount rate used, consideration should be given to sensitivity analyses that use differential rates for costs and outcomes and/or that vary the rate between 0% and 6%[3].