Assessing the extent of drug interactions among patients with multimorbidity in primary and secondary care in the West Midlands (UK): a study protocol for the Mixed Methods Multimorbidity Study (MiMMS)

Abstract

IntroductionThe numbers of patients with three or more chronic conditions (multimorbidity) are increasing, and will rise to 2.9 million by 2018 in the UK alone. Currently in the UK, conditions are mainly managed using over 250 sets of single-condition guidance, which has the potential to generate conflicting recommendations for lifestyle and concurrent medication for individual patients with more than one condition. To address some of these issues, we are developing a new computer-based tool to help manage these patients more effectively. For this tool to be applicable and relevant to current practice, we must first better understand how existing patients with multimorbidity are being managed, particularly relating to concerns over prescribing and potential polypharmacy.Methods and analysisUp to four secondary care centres, two community pharmacies and between four and eight primary care centres in the West Midlands will be recruited. Interviewees will be purposively sampled from these sites, up to a maximum of 30. In this mixed methods study, we will perform a dual framework analysis on the qualitative data; the first analysis will use the Theoretical Domains Framework to assess barriers and enablers for healthcare professionals around the management of multimorbid patients; the second analysis will use Normalisation Process Theory to understand how interventions are currently being successfully implemented in both settings. We will also extract quantitative anonymised patient data from primary care to determine the extent of polypharmacy currently present for patients with multimorbidity in the West Midlands.DiscussionWe aim to combine these data so that we can build a useful, fully implementable tool which addresses the barriers most amenable to change within both primary and secondary care contexts.Ethics and disseminationFavourable ethical approval has been granted by The University of Birmingham Research Ethics Committee (ERN_16–0074) on 17 May 2016. Our work will be disseminated through peer-reviewed literature, trade journals and conferences. We will also use the dedicated web page hosted by the University to serve as a central point of contact and as a repository of our findings. We aim to produce a minimum of three articles from this work to contribute to the international scientific literature.Protocol registration numberNIHR Clinical Research Network Portfolio Registration CPMS ID 30613.

A Process Mining and Text Analysis Approach to Analyse the Extent of Polypharmacy in Medical Prescribing

“…We focus on text processing to transform unstructured clinical records into suitable data for process mining; a process mining, generalisation and visualisation approach appropriate for the medication prescription data; and obtaining initial results on polypharmacy (Section IV). We found that almost 62% of patients in our sample received medication with some level of interaction, which our clinical study [5] will explore in more detail. We provide initial analysis in Section V, and briefly discuss some of the problems we encountered in obtaining and processing the data, which could be significantly mitigated by relatively simple changes and adopting a mindset of curating data for the purposes of automated analysis.…”

“…The analysis showed the feasibility of applying a process mining approach to this type of data, despite the concept of a 'prescription process' being loosely defined and differing from the usual conception of a process, for example having a start and end point defined only by the parameters of the data extract. The resulting models exhibit a variety of processlike structures, and the initial polypharmacy results are of great interest to begin to answer the questions raised by the protocol [5]. Detailed analysis will now be carried out, to relate polypharmacy to different patient cohorts, sites, clinical conditions, and to investigate root causes.…”

“…A grammar was developed to parse the unstructured 'Dose' to extract quantity and time information 5 . The basic form of the field is Dose ::-<directive> <quantity> <unit> <frequency> <time-spec> as seen in the examples in Table II.…”

“…Polypharmacy refers to the concurrent prescription of multiple medications for a single patient [18], [22]. It is a growing problem [5], in particular due to increasing numbers of patients suffering from multiple chronic diseases (multimorbidity), arising from factors including ageing populations and improved treatment of chronic disease [46]. The increased burden of treatment is a particular concern for those with multimorbidity as they frequently take many medications related to each individual condition.…”

“…Secondly, what frequentlyoccurring patterns of prescription can be identified to give insight into the underlying causes of polypharmacy and enable mitigations to be developed? More details of the clinical rationale can be found in the study protocol [5].…”