Get started
Work

Linking data to explore variation in care home prescribing: a helpful tool for starting conversations to enable change

Year: 2025
Type: editorial
Abstract: Editorial to accompany Antibiotic prescribing for care-home residents: a population-based, cross-classified multilevel analysis in Scotland, UK[3] Cross-sectoral data linkage offers an inclusive way to explore everyday practice at scale. It is time-consuming to link new data sources and meaningful individual characteristics are not routinely captured in secondary care data. Prescribing for residents is rarely done by care homes, thus interventions to improve antimicrobial use must also engage primary care prescribers. Care home prescribing data can be a useful tool to start conversations with key stakeholders to collaborate and create change to benefit residents. Following advances achieved linking routinely-collected health data, there has been growing interest in the potential of cross-sectoral data linkage to generate evidence to improve care for older people [1]. Social care data sources are a particularly precious resource, under-utilised in the UK to-date [2]. Reusing existing de-identified data to enable inclusion of adults living in care homes in research, represents progress for a population commonly excluded and poorly understood. As such, we welcome the innovative research by De Souza et al. which joined individual-level care home resident data to community prescribing data, using care home characteristics and general practice characteristics to explore variation in antibiotic prescribing [3]. Medicines safety and appropriate prescribing are topics of interest for this population, who are particularly vulnerable to adverse effects and inappropriate polypharmacy [4]. They found variation in prescribing practice, with some association with resident characteristics [3]. However, there was significant unexplained variation between care homes and general practices (GPs), signalling the need to involve both in prescribing-focused initiatives [3]. This echoes previous analysis which looked at variation in antibiotic prescribing across care homes elsewhere in the UK [5]. Variation has also been observed in the dispensing of psychotropic and anticholinergic medications for people living in care homes, not explained by differences in care home quality measures [6]. A challenge faced by these studies are the relative lack of meaningful variables about residents themselves. In addition to age and sex the team included dementia (based on prescribing or hospital discharge diagnosis), with a recorded prevalence of 46%. They derived a Charlson Comorbidity Index from hospital discharge data (excluding dementia) and found 63.5% had no comorbidities. While this may reflect the utility/limitations of the measure, these findings reflect important biases in the data sources and the threshold of a secondary care diagnosis necessitating a hospitalisation and comprehensive diagnostic coding from discharge correspondence. Improved accessibility of primary care data for linkage would enable a more inclusive assessment of long-term conditions important in the population, which do not necessarily result in hospitalisation [7]. In addition, linkage involving more granular data collected by care home staff, using existing initiatives such as the Scottish Care Home Census [8] or through resident Digital Care Record software [9] has enabled more individual resident variables to be incorporated, enriching insights about the population. It is also important to acknowledge that detecting and managing infections among people living in care homes is complex. The global context of concern around antimicrobial resistance is set alongside professional and public concern around under-recognition of infection and fatality associated with sepsis. Infections, most commonly involving the respiratory system or urinary tract, remain common causes of death among the care home population [10]. Infections also feature on lists of ambulatory-care sensitive conditions as ‘Acute conditions that could potentially have been avoided if the patient had been better managed in primary care’ [11], diagnoses in which hospitalisation is deemed unnecessary. Consequently, care staff face challenges in ensuring timely and appropriate treatment for those they care for. It is rare in the UK for care homes to directly employ prescribers, so access to prescribed treatments rely on other professionals, primarily advanced nurse practitioners and GPs working outside the care home. There are a broad range of contractual arrangements and services in place to provide primary care support to care homes, including scheduled in-person sessions on fixed days, resources for proactive care management or more reactive responses, akin to those delivered to the general population. However, these specific models of primary care were not included in the analyses by De Souza et al. and are typically not operationalised in modelling of care home prescribing and resource use. This is a limitation given the increasing pressures on primary care and varied support care homes can draw upon. An indirect measure which De Souza et al. do include is the number of GP practices which a care home has residents registered with. In their sample, nearly a third of homes had residents registered with more than eight GPs. Where previously this may have reflected individual preference, retaining registration with existing GP, increasingly this is a marker of a system in crisis, with new residents registered on any practice list with capacity, with decreasing likelihood of dedicated care home time for the practitioners and increased complexity and burden for care home staff seeking support. Sampling variation (predominantly urine) was highlighted by the team as associated with higher rates of antimicrobial prescribing [3]. However, the practicalities of this at a care home level are not insignificant. This includes barriers to sampling when primary care is closed and obtaining samples where a significant proportion of residents are living with incontinence. Including recognition and prevention of infection into fundamental proactive resident care has been demonstrated as effective in supporting changes in practices in this area, rather than educational approaches in isolation [12]. The pivotal contribution of involving and empowering care home staff, drawing on their expertise and relationships with residents have been key in changing use of psychotropic medication use across homes [13]. This work also reinforced the inter-dependencies of care homes and their external prescribers in effecting change for residents and the variation in response of those prescribers when met with requests to reduce or stop medications [13]. Noting the gap between the age of the data included (2016–17) and time of publication may cause some to question their utility and relevance. In our experience, this reflects the under-appreciated complexity associated with linking care home data to other data sources [14] and we particularly commend the researchers for the preparatory work required which has enabled analyses at the care home level [3]. The workforce pressures on UK primary care and changes in delivery of care post-pandemic are likely to have exacerbated, rather than reduced variation. We welcome the Vivaldi Social Care programme, which is enabling routine sentinel surveillance of infections across care homes in England, though a routine data linkage approach, to enable a data-informed approach to infection management going forwards [15]. We look forward to further insights from the Antibiotic Research in Care Homes team’s research programme. We support the role of improving accessibility and utility of routine data on care home prescribing to start conversations with key stakeholders, facilitating collaboration to create change to benefit residents. No conflicts of interest to declare. J.K.B. is supported by an NHS Education for Scotland/Chief Scientist Office Postdoctoral Clinical Lectureship (PCL/21/01). The funders played no role in the writing of this editorial.
DOI: 10.1093/ageing/afaf061
Language: English
Open Access status: bronze
OpenAlex ID: W4408505105
Cites: 14
Cited by:
Related to: 10