Can timely recognition of deteriorating health prevent hospitalisation of older community-dwelling adults? The effect of a notification algorithm and a digital decision support tool

Research output: ThesisPh.D. thesis


The challenges imposed by an ageing demography will confront healthcare systemsin the years to come. Even highly effective health care systems will struggle to meetthe demands of ageing populations as higher age is associated with multimorbidity,functional decline, and increased use of healthcare services.

In recent years, comprehensive national healthcare reforms have been implemented in most of the Nordic countries, centralising the secondary healthcare sector and introducing more specialised treatment and care in the primary healthcare sector. This has led to increased political focus on improving the collaboration and communication between healthcare sectors, especially regarding older community-dwelling adults. The goal is to improve continuity of care and patient satisfaction through prevention
of hospital admissions and continued independent living in the community. A common characteristic of these policies is the development of a more advanced level of home nursing practice to support early recognition and timely treatment of emerging disease in older community-dwelling adults.

Recognition of emerging disease in older community-dwelling adults can be difficult, due to atypical symptoms and complex health issues. Algorithms and prediction models integrated into digital solutions might assist home care nurses in recognising early health deterioration. We have previously shown that home care services increase prior to acute short-term admission in older community-dwelling adults. Based on this knowledge, we built the Prevention of AcuTe admIssioN Algorithm (PATINA), a novel
algorithm and decision support tool that analyses administrative data on home care utilisation and notifies home care nurses when home care services increase.

The purpose of this thesis was: 1) to examine the efficacy of the PATINA tool on health care utilisation of older community-dwelling adults when implemented as an aid to home care nurses in three Danish municipalities, 2) to translate and culturally adapt into Danish a frailty measure with wide transdisciplinary acceptance and validated applicability and to test its reliability for key health professionals in the crosssectoral collaboration, 3) to examine the association between frailty and six-month all-cause mortality, care home admittance, hospitalisation and temporary care in a skilled nursing facility in older community-dwelling adults, and 4) to develop a framework supporting the development, implementation, evaluation, and monitoring of acute community home care services.

In a protocol paper (Paper 1), we described an open-label stepped-wedge clusterrandomised controlled trial to evaluate the efficacy of the PATINA tool when implemented as an aid to home care nurses in three Danish municipalities of varying sizeand organisation. Primary outcome was the number of hospital admissions, i.e., within 30 days of being identified by the PATINA tool as ‘at risk’ of hospitalisation.The secondary outcomes were hospital readmissions, other hospital admissions, outpatient contacts, primary care physician contacts, use of temporary care, and mortality, all within 30 days of identification.

The PATINA tool consists of two modules: an algorithm and a decision support tool.The algorithm monitors and analyses the use of municipal home care services bysummarising the time delivered for providing personal care and domestic help, training or rehabilitation, and nursing care. On a weekly basis, home care nurses werenotified of older adults with increased use of home care services >17.5% during thelast month compared to a period of 6 months prior, which we considered to predict anincreased risk of adverse health outcomes. For each notified citizen, the home carenurse received a graph visualising the citizen’s use of home care for the past sixmonths as well as a link to access the PATINA decision support tool. This tool wasdesigned to help nurses reflect upon subtle signs of health deterioration while assessing the citizen’s health situation using validated health assessment scales andinformation on risk factors of acute disease. However, decisions on further actionwere at the nurse’s discretion.

As part of the preparation of the PATINA decision tool, we translated and culturallyadapted the Clinical Frailty Scale (CFS) into the Danish language (Paper 2 and 3).We added to the potential use of the CFS in the cross-sectoral collaboration by displaying its excellent inter-rater reliability both within and across four groups of healthprofessionals (community nurses, primary care physicians, intensive care physicians,and hospital physicians from internal medicine). The inter-rater reliability was high inall four professional groups, ranging from ICC 0.81 to 0.90, indicating the potential fora common language on frailty. 

Based on a stepped-wedge cluster-randomised controlled trial design, the PATINAtool was sequentially implemented over a period of 12 months from 1 June 2020 to31 May 2021 in the area home care teams of each of the three municipalities (Paper 4). In total, 2,464 participants were included in the study - 1,216 (49.4%) in the control phase and 1,248 (50.6%) in the intervention phase. In the control phase, 102 individuals were hospitalised ≤30 days of notification by the PATINA algorithm in 33,943days of risk, compared with 118 admissions ≤30 days in 34,843 days of risk in the intervention phase. The PATINA tool was not associated with a reduction in first hospitalisation ≤30 days after notification (IRR 1.10, 95%CI 0.87-1.50, p= 0.38), but with a59% reduction in readmissions ≤30 days (IRR 0.41, 95%CI 0.24-0.68, p= <0.001), a140% increase in contacts to primary care physicians (IRR 2.40, 95%CI 1.18-3.20, p<0.001) and a 150 % increase in use of temporary care in a skilled nursing facility.However, the effect differed across the three municipalities. This was likely explainedby differences in capacity (e.g., organisational structure and size) and capability (e.g.,staff composition and experience). Our results suggest that the PATINA tool shiftedhealth care tasks from secondary to primary care.

Frailty has gained attention for its potential to identify older adults at risk of adversehealth outcomes. However, little is known about the predictive value of frailty in adultscharacterised by increasing need of home care. In a pre-planned prospective cohortanalysis of the intervention phase of the PATINA study (Paper 5), we examined theassociation between frailty and six-month all-cause mortality, hospitalisation, temporary care in a skilled nursing facility, and institutionalisation into a care home. Wefound frailty to be significantly associated with higher risk of 6-month all-cause mortality, admission to hospital, and temporary care, but not with care home. Surprisingly, risks did not increase proportionally with mounting level of frailty. 

All Danish municipalities have implemented an acute community healthcare serviceto support continuity of care by acting as a bridge between the primary and secondary healthcare sectors in the early recognition of acute disease and provision of treatment and care. Despite similar political priorities in many countries, however, the literature on the organisation and effect of such services is limited. We present a conceptual framework that describes acute community healthcare services (Paper 6) andcan be used for understanding the concept as well as supporting the developmentand comparison of these services in other healthcare organisations. For illustrativepurposes, we apply the framework to the Danish acute community healthcare services using implementation data from 2020 and identify gaps and opportunities forlearning. The framework identifies two key pairs of dimensions: (1) capacity & capability, and (2) coordination & collaboration. These dimensions, together with the governance structure and quality assurance initiatives, are of key importance to the effect of acute community healthcare services. Application of the framework indicatedconsiderable variation in the approaches of the Danish municipalities.
Translated title of the contributionKan tidlig opsporing af svigtende helbred forebygge indlæggelse af ældre hjemmeplejemodtagere? : Effekten af en notifikationsalgoritme og et digitalt beslutningsstøtteværktøj
Original languageEnglish
Awarding Institution
  • University of Southern Denmark
  • Andersen-Ranberg, Karen, Principal supervisor
  • Espersen, Kurt, Supervisor, External person
  • Bech, Mickael, Supervisor
  • Lauridsen, Jørgen T, Supervisor
  • Wiil, Uffe Kock, Supervisor
  • Kidholm, Kristian, Supervisor
External participants
Date of defence28. Apr 2023
Publication statusPublished - 21. Apr 2023

Note re. dissertation

Print copy of the thesis is restricted to reference use in the Library. 


  • Algorithm
  • Decision support tool
  • Frailty
  • Acute community health care services
  • Older community-dwelling adults
  • Adverse health outcome


Dive into the research topics of 'Can timely recognition of deteriorating health prevent hospitalisation of older community-dwelling adults? The effect of a notification algorithm and a digital decision support tool'. Together they form a unique fingerprint.

Cite this