Skip to main content
Download PDF

Reasons for hospitalisation and cumulative mortality in people, 75 years or older, at high risk of hospital admission: a prospective study

BMC Geriatrics volume 24, Article number: 176 (2024) Cite this article

Abstract

Background

A small proportion of the older population accounts for a high proportion of healthcare use. For effective use of limited healthcare resources, it is important to identify the group with greatest needs. The aim of this study was to explore frequency and reason for hospitalisation and cumulative mortality, in an older population at predicted high risk of hospital admission, and to assess if a prediction model can be used to identify individuals with the greatest healthcare needs. Furthermore, discharge diagnoses were explored to investigate if they can be used as basis for specific interventions in the high-risk group.

Methods

All residents, 75 years or older, living in Östergötland, Sweden, on January 1st, 2017, were included. Healthcare data from 2016 was gathered and used by a validated prediction model to create risk scores for hospital admission. The population was then divided into groups by percentiles of risk. Using healthcare data from 2017–2018, two-year cumulative incidence of hospitalisation was analysed using Gray´s test. Cumulative mortality was analysed with the Kaplan–Meier method and primary discharge diagnoses were analysed with standardised residuals.

Results

Forty thousand six hundred eighteen individuals were identified (mean age 82 years, 57.8% women). The cumulative incidence of hospitalisation increased with increasing risk of hospital admission (24% for percentiles < 60 to 66% for percentiles 95–100). The cumulative mortality also increased with increasing risk (7% for percentiles < 60 to 43% for percentiles 95–100). The most frequent primary discharge diagnoses for the population were heart diseases, respiratory infections, and hip injuries. The incidence was significantly higher for heart diseases and respiratory infections and significantly lower for hip injuries, for the population with the highest risk of hospital admission (percentiles 85–100).

Conclusions

Individuals 75 years or older, with high risk of hospital admission, were demonstrated to have considerable higher cumulative mortality as well as incidence of hospitalisation. The results support the use of the prediction model to direct resources towards individuals with highest risk scores, and thus, likely the greatest care needs. There were only small differences in discharge diagnoses between the risk groups, indicating that interventions to reduce hospitalisations should be personalised.

Trial registration

clinicaltrials.gov Identifier: NCT03180606, first posted 08/06/2017.

Peer Review reports

Background

A small proportion of the older population accounts for a high proportion of healthcare use [1,2,3,4]. For effective use of limited healthcare resources, it is of great importance to identify the group with the greatest needs. Individuals with multimorbidity and/or frailty are more likely to be admitted to hospital [5, 6]. In addition, frail individuals have a higher mortality rate than others [7]. The prevalence of both multimorbidity and frailty increases with age [8,9,10]. People 65 years or older have a hospitalisation prevalence of 5.6 to 25.8 percent over 12 months [11]. Data from previous studies describing diagnostic reasons for hospitalisation in older people is scarce. Older people seem, however, to be hospitalised most frequently due to diseases in the circulatory and respiratory system [12, 13]. Men seem to be hospitalised more frequently due to cancer and diseases in the respiratory system than women, while women seem to be hospitalised more frequently due to fractures or falls than men [12].

In Sweden, primary care consists of more than 1 000 health centres and clinics, which are the first line of contact to the healthcare system for the whole population. These health centres are responsible for cooperating with other levels of care as well as the municipalities, which are responsible for long term care such as nursing homes and domestics services [14].

The world is facing a situation with an increasing number of older people [15, 16]. Hence, it is important to have methods to direct the resources to the individuals with the highest risk of hospital admission when planning healthcare interventions preventing decline in health [17, 18]. Different ways of predicting patients with high healthcare use have been suggested, including frailty [7, 19, 20] and multimorbidity [6, 10]. Many strategies for predicting high-risk patients rely on clinical assessments or self-administrated questionnaires. These screening procedures are costly, and some patients at risk will go unnoticed due to lower response rate in groups with characteristics associated with risk of high healthcare use [21]. The use of healthcare databases to predict who will be admitted to hospital has been suggested to be a more feasible approach [22,23,24,25,26,27,28]. To focus on prevention of hospitalisation is not only relevant from a cost perspective; hospitalisation is also associated with higher risk of admission to nursing home [29] and functional decline, which negatively impacts the quality of life of the individual [30].

To identify individuals with a probable benefit from interventional programs, Marcusson et al. [31] designed a model aiming to predict hospital admission among people, 75 years or older, in the coming twelve months. The prediction model was created to be used in a pragmatic intervention trial in primary care, called Proactive healthcare for frail elderly persons. In this trial, individuals identified by the model to have a high risk of hospital admission, were assessed with comprehensive geriatric assessment (CGA) adapted to primary care [32]. The aim was to investigate if personalised, targeted primary care could reduce hospitalisations compared with usual care. The assessment of identified people, 75 years or older, resulted in a 22 percent relative risk reduction for inpatient-hospital days [33]. The intervention was cost-effective [34].

The knowledge about primary discharge diagnoses and how they differ between individuals at different levels of risk of hospitalisation is scarce. An identifiable difference in distribution of diagnoses could form the basis for targeted intervention efforts in different risk groups of frail old people. This can be of value, for example in primary care, in efforts to organise cost-effective care for large populations and to proactively reach older individuals at high risk of morbidity, hospitalisation, and increased risk of mortality.

The present study examines how the cumulative incidence of hospitalisation, the reason for hospitalisation, and the cumulative mortality varies at different levels of risk of hospital admission, in a Swedish population at an age of 75 years or older, to further validate the prediction model developed by Marcusson et al. [31].

Aim

The aim of this study was to explore frequency and reason for hospitalisation and cumulative mortality, in an older population at predicted high risk of hospital admission and to assess if a prediction model can be used to identify individuals with the greatest healthcare needs. The aim was further to examine if differences in distribution of primary and secondary discharge diagnoses could form the basis for targeted intervention efforts according to gender, or to risk groups of older people.

Methods

Setting and population

The study is a prospective registry-based cohort study that included all residents, 75 years or older on January 1st, 2017, in the county of Östergötland in the southeast part of Sweden. Healthcare data consisting of gender, age, and diagnoses from both hospital care and open clinic visits (grouped by two digits) according to International Classification of Diseases (ICD-10) [35] from the year of 2016, was gathered from the Care Data Warehouse (CDW) of Region Östergötland. The CDW is a computerised information system where healthcare use for the region is stored. The number of hospital admissions (at the latest initiated on December 31st, 2016) and the number of non-physician (for example nurses, occupational therapists, or physiotherapists), physician, and emergency room (ER) visits were also included. In total, 37 variables were used in a prediction model to create a risk score for hospital admission for each person [31]. The number of hospital admissions (initiated between January 1st, 2017, and December 31st, 2018), the related primary and secondary discharge diagnoses, and the mortality during 2017 and 2018 were also collected from the CDW. If an individual died during hospitalisation this was regarded both as a hospital admission episode and as a death. In those cases, the first registered diagnosis was regarded as the discharge diagnosis.

The study population was divided into risk groups by percentiles of risk, derived from the predicted risk of hospital admission. Individuals with risk percentiles < 60 were regarded as one group as these individuals were considered less clinically relevant when primarily screening for high-risk patients, which is the main purpose of the prediction model [31]. This group was therefore not further analysed. Individuals with risk percentiles 60 to 100 was divided in groups of five percentile points.

Statistics

The cumulative incidence of hospitalisation in the different risk groups was analysed using Gray´s test. The subdistribution hazard ratios (SHR) were analysed with Fine-Gray subdistribution hazard model, treating mortality as competing risk [36, 37]. The cumulative mortality rates were calculated with the Kaplan–Meier method, and the hazard ratios were analysed with Cox-regression. The analyses were also conducted according to gender, and confidence intervals were compared to find any significant difference between gender.

Each individual’s primary discharge diagnosis of their first episode of hospitalisation were analysed, both as blocks, according to ICD-10 [38], and as ICD-10 diagnoses (grouped by two digits). ICD-10 blocks were sorted by the most frequent to the least frequent. ICD-10 blocks accounting for at least two percent of the primary discharge diagnoses for the population in total were analysed separately. The remaining ICD-10 blocks were clustered into one group called other diagnoses. Based on the results from the analysis on hospitalisation, where the cumulative incidence overlapped up to percentile 85, the ICD-10 blocks were analysed in five groups: percentiles < 60, 60–85, 85–90, 90–95, and 95–100 and in three groups: percentiles < 60, 60–85, and 85–100. The distribution of the top ten ICD-10 diagnoses for the entire population were presented in total and according to gender. The primary discharge blocks and diagnoses were analysed with Chi-2 and standardised residuals (Pearson residuals) to find significant differences between risk groups and gender. Each individual´s secondary discharge diagnoses (ICD-10 diagnoses grouped by two digits) of their first episode of hospitalisation were analysed with Chi-2 and standardised residuals. Top five secondary discharge diagnoses were presented. Statistical analyses were performed in R v.4.2.1. Gray and Fine-Gray analyses were computed in the cmprsk-package.

Results

In total, 40 618 individuals, 75 years or older, living in Östergötland, Sweden, on January 1st, 2017, were identified. The background characteristics of the population are reported in Table 1. Mean age was 82 years, and 57.8 percent of the population were women. The percentage of women decreased with increasing risk for hospitalisation in the groups.

Table 1 Characteristics of the study population at baseline and their healthcare use during 2016
Full size table

In total, 8 131 individuals had experienced at least one episode of hospitalisation after one year and 13 491 after two years. The individuals in the higher risk groups had higher cumulative incidence of hospitalisation. The cumulative incidence, during a follow-up period of two years, varied between 24 percent in the lowest (percentiles < 60) compared to 66 percent in the highest risk group (percentiles 95–100). The confidence intervals overlapped between adjacent risk groups from percentiles 60–65 up to percentiles 80–85 and were thereafter separated between the highest risk groups (percentiles 85–90, 90–95, and 95–100). There was no significant difference in hospitalisation between men and women when divided into risk groups. Data on cumulative incidence and SHR are presented in Table 2 and the cumulative incidence is also presented in Fig. 1.

Table 2 Number of hospital admissions, 2-year cumulative incidence of hospitalisation and subdistribution hazard ratios, year 2017–2018
Full size table
Fig. 1
figure 1

2-year cumulative incidence of hospitalisation by percentiles of risk of hospital admission, year 2017–2018. Estimated using Gray´s test

Full size image

With a follow-up time of two years, the mean inpatient-hospital days were 3.6 days, ranging from 2.2 days in the lowest risk group (percentiles < 60) to 10.6 days in the highest risk group (percentiles 95–100). The mean number of days during the first hospitalisation episode was 6.2, ranging from 6.0 (percentiles < 60) to 6.4 (percentiles 95–100).

The cumulative mortality varied between 7 percent (percentiles < 60) and 43 percent (percentiles 95–100) with a follow-up time of two years. The confidence intervals overlapped between adjacent risk groups from percentiles 60–65 up to percentiles 85–90 and were thereafter separated (percentiles 90–95 and 95–100), see Table 3. Data on mortality are also presented in Fig. 2. There was no significant difference between men and women when divided into risk groups, based on overlapping confidence intervals.

Table 3 Cumulative mortality by percentiles of risk of hospital admission, year 2017–2018
Full size table
Fig. 2
figure 2

2-year cumulative mortality by percentiles of risk of hospital admission, year 2017–2018. Estimated using Kaplan–Meier

Full size image

In total there were 476 different primary discharge ICD-10 diagnoses after one year and 561 after two years. Ten ICD-10 diagnoses independently constituted 2.0 percent or more of all hospitalisation episodes. The diagnoses are listed in Fig. 3. When analysing primary discharge diagnosis according to ICD-10 blocks, there were significant differences between the groups (percentiles < 60, 60–85, 85–90, 90–95, and 95–100). However, the differences in percentage units were small and of unclear clinical relevance, and therefore only results divided into three risk groups are presented. The most common reason for hospitalisation according to ICD-10 blocks, divided into three risk groups are listed in Table 4. Other forms of heart disease (I30-I52) was the most common reason for hospitalisation, both in the study population in total and in all the risk groups. The second most frequent reason for all people, 75 years or older, and in the highest risk groups (percentiles 85–100) was influenza and pneumonia (J09-J18). Influenza and pneumonia was a frequent reason for hospitalisation in the lowest two risk groups (percentiles < 60 and 60–85) as well, but injuries to the hip and thigh (S70-S79) was more frequent.

Fig. 3
figure 3

Distribution of top 10 primary discharge diagnoses for men and women, year 2017–2018. Presented according to ICD-10. S72 Fracture of femur, I63 Cerebral infarction, I50 Heart failure, N39 Other disorders of urinary system, I21 Acute myocardial infarction, J18 Pneumonia, organism unspecified, I48 Atrial fibrillation and flutter, J15 Bacterial pneumonia, not elsewhere classified, R07 Pain in throat and chest, J44 Other chronic obstructive pulmonary disease. *Statistical significance (Chi-2 test, p < 0.05) between men and women

Full size image
Table 4 Distribution of primary discharge diagnosis according to ICD-10 block by risk of hospital admission
Full size table

Among the ICD-10 diagnoses for the study population in total, fracture of femur (S72) was the most frequent reason for hospitalisation (6.3%), cerebral infarction (I63) the second (4.3%), and heart failure (I50) the third (3.8%) most frequent reason. For women, 75 years or older, fracture of femur was the most frequent discharge diagnosis (8.0%), cerebral infarction the second (4.3%), and atrial fibrillation and flutter (I48) the third most frequent (3.6%). For men, 75 years or older, the most frequent discharge diagnosis was heart failure (4.8%), the second acute myocardial infarction (I21) (4.5%), and the third other disorders of urinary system (N39) (4.4%).

For other chronic obstructive pulmonary disease (J44) 43 percent of the patients had the same discharge diagnosis the second episode, for the diagnosis heart failure 31 percent, and for atrial fibrillation and flutter 29 percent. These were the three diagnoses with the highest number of readmissions.

The five most frequent secondary discharge ICD-10 diagnoses were essential hypertension (I10) (12.1%), atrial fibrillation and flutter (I48) (6.1%), chronic ischaemic heart disease (I25) (4.4%), personal history of medical treatment (Z92) (4.2%), and type 2 diabetes mellitus (E11) (3.9%). For the highest risk group (percentiles 85–100) essential hypertension had a significant lower frequency than expected. For atrial fibrillation and flutter, chronic ischaemic heart disease, and personal history of medical treatment the highest risk group had a significant higher frequency than expected. For type 2 diabetes mellitus no significant difference from expected was found.

Discussion

In the present study, the cumulative incidence of hospitalisation, the reasons for hospitalisation by diagnoses, and the cumulative mortality was explored to investigate how it varied between individuals, 75 years or older, at different levels of predicted risk of hospital admission.

The cumulative incidence of hospitalisation and the cumulative mortality increased with increasing risk of hospital admission, regardless of gender. From percentile 80 and upwards, the SHR for hospital admission ranged from 2.3 to 4.1 compared with individuals in the percentile < 60. The hazard ratios for mortality in the same percentiles ranged from 3.8 to 7.5.

It is well established that frail people have both higher mortality rate and higher risk of hospitalisation [6, 10, 39]. Individuals with high risk of hospital admission, according to the prediction model, also had higher risk for these outcomes. The increased risk for the population over percentile 80 was in the same range as frail individuals selected by the electronic frailty index developed by Clegg et al. [40]. However, in Clegg´s study the population was younger than in the present study, and the hazard ratios was calculated comparing frail individuals with fit individuals. The percentile < 60 has not been analysed in the present study, and it can therefore not be established whether these individuals were fit or not. Furthermore, frailty is more common among women than men [7, 19, 41], but the proportion of women decreased with increasing risk of hospitalisation in this study. Hence, even if frailty overlaps with risk of hospital admission, as expressed with the prediction model, they can be regarded as complimentary measures of risk, or vulnerability in an older person. We believe that the model could be used to identify vulnerable adults, 75 years or older, as a possible first step in screening for frailty.

As the frail population is heterogenous regarding care needs, the prediction model may have the advantage of finding individuals with the greatest needs. On the other hand, frail individuals that do not have frequent contacts with healthcare will not be identified by the model. Additional clinical methods will therefore always be of importance. To provide significant patient benefits, the use of the prediction model should be combined with a holistic clinical assessment of the patient. Future studies are needed to better understand the differences and similarities between frailty and predicted risk of hospital admission. It would also be valuable to investigate if the model can predict other negative outcomes, for example nursing home admission, or loss of functional ability.

Individuals with the highest risk of hospital admission were expected to have longer average length of stay during their first hospitalisation episode, reflecting complex medical needs and/or frailty. Several studies, but not all [42], have shown a correlation between frailty and average length of hospital stay [41, 43,44,45]. In the present study, the mean number of days for a hospitalisation episode did not differ between the risk-groups. However, the mean number of in-hospital days during the 2-year follow-up was considerably higher in the group with the highest risk of hospital admission.

The most common reason for hospitalisation were related to the circulatory and respiratory system, agreeing with previous studies [12, 13]. Women, 75 years or older, were hospitalised for femur fractures as the most frequent ICD-10 diagnosis, while for men, 75 years or older, heart failure was the main reason for hospitalisation.

The primary discharge diagnoses according to ICD-10 blocks are presented in three risk groups (percentiles < 60, 60–85, and 85–100) in this study. First, the ICD-10 blocks were analysed divided in five risk groups (percentiles < 60, 60–85, 85–90, 90–95, and 95–100). Even though several significant values could be found, the differences in percentage units were minimal. To make the result more manageable, the discharge diagnoses were presented in three risk groups, considering that the confidence intervals for the cumulative incidence of hospitalisation overlapped between adjacent risk groups between percentile 60 to 85. With this taken into consideration, the population, 75 years or older, could be divided into three groups (percentile < 60, 60–85, and 85–100). Even when presented in three risk-groups the differences were small and of unclear clinical relevance.

Pneumonia, chronic obstructive pulmonary disease (COPD), and heart failure are regarded as diagnoses where hospital admissions can be prevented if the patient is given optimal care in primary care and municipality [46,47,48]. These diagnoses were significantly more common in the highest risk-group (percentiles 85–100) in this study. COPD and heart failure were also the most frequent reasons for readmission with the same diagnosis as in the first hospitalisation episode. For women, 75 years or older, femur fracture was the most frequent discharge diagnosis. Thus, targeting risk for falls and pneumonia, as well as suboptimal treatment of COPD and heart failure could be part of an intervention to lower the rate of hospitalisation in a high-risk group. However, the difference in percent units was small, indicating that interventions to reduce hospitalisations should be personalised.

One strength of the study is that all individuals, 75 years or older, who lived in the county of Östergötland during the study period were included in the analysis. All hospitalisation episodes were included, regardless of where they occurred in Sweden. The registry used in the present study has almost non-existent dropouts and very little (negligible) missing data. There is, however, a small risk that some individuals moved from the region, which could impact the analyses of the hospitalisation, discharge diagnoses, and cumulative mortality.

Sweden has a low number of hospital beds per 1 000 people (2.1) compared with the mean for the European union (4.6) and even compared with the world in total (2.9) [49]. The average length of hospital stay in Sweden 2018 was 5.5 days, the fourth shortest in the European Union [50]. The relatively low number of hospital beds and short length of stay likely affect the average length of stay, the cumulative incidence of hospitalisation, and which diagnoses caused the hospitalisation. This makes it necessary to externally validate the model in other countries to be able to draw conclusions on the prediction model´s performance in other healthcare systems and generalisation of results. Further, discrepancies in coding of ICD-10 diagnoses between healthcare units and changes in the coding practice may impact the internal validity of the model. In order to avoid misclassification of individuals into high-risk or low-risk, it is of importance to combine the prediction model with a clinical assessment.

Conclusion

Individuals, 75 years or older, identified by the prediction model to have high risk of hospital admission were demonstrated to have considerably higher cumulative incidence of hospitalisation, as well as mortality than individuals with lower risk. The results support the use of the prediction model to direct interventions towards groups in a population, 75 years or older, with the highest risk scores, and thus, likely the greatest care needs. In order to work resource-efficiently and reduce the risk of inpatient care, it is important to proactively identify these adults at high risk. The population selected by the prediction model partly overlap the population with frailty, but more studies are needed to describe this relation in more detail.

Heart failure, COPD, and pneumonia were significantly more frequent reason for hospitalisation for the individuals in the highest risk group. The women, 75 years or older, had fractures of the femur, and the men, 75 years or older, had heart failure as the most frequent discharge diagnosis. There were only small differences in discharge diagnoses between the risk groups, indicating that care interventions to reduce hospitalisations should be personalised and holistic, rather than disease-specific for people, 75 years or older, at high risk.

Availability of data and materials

The dataset used and/or analysed during the current study is available from the corresponding author on reasonable request.

Abbreviations

CDW:

Care Data Warehouse

CGA:

Comprehensive Geriatric assessment

CI:

Confidence interval

COPD:

Chronic obstructive pulmonary disease

ER:

Emergency room

ICD-10:

International Classification of Diseases 10th Revision

SD:

Standard deviation

SHR:

Subdistribution hazard ratio

References

  1. Schoenman JA, Chockley N. The concentration of health care spending. National Institute for Health Care Management Research Educational Foundation; 2012. https://nihcm.org/assets/articles/databrief3final.pdf. Cited 2023 March 22.

  2. Holle M, Wolff T, Herant M. Trends in the concentration and distribution of Health Care expenditures in the US, 2001–2018. JAMA Netw open. 2021;4(9):e2125179-e.

    Article  Google Scholar 

  3. Tanke MA, Feyman Y, Bernal-Delgado E, Deeny SR, Imanaka Y, Jeurissen P, et al. A challenge to all. A primer on inter-country differences of high-need, high-cost patients. PLoS One. 2019;14(6):e0217353.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  4. Cohen SB, Yu W. Statistical brief# 354: the concentration and persistence in the level of health expenditures over time: estimates for the US population, 2008–2009. Rockville: Agency for Healthcare Research and Quality; 2012.

    Google Scholar 

  5. Nägga K, Dong H-J, Marcusson J, Skoglund SO, Wressle E. Health-related factors associated with hospitalization for old people: comparisons of elderly aged 85 in a population cohort study. Arch Gerontol Geriatr. 2012;54(2):391–7.

    Article  PubMed  Google Scholar 

  6. Ilinca S, Calciolari S. The patterns of health care utilization by elderly europeans: frailty and its implications for health systems. Health Serv Res. 2015;50(1):305–20.

    Article  PubMed  Google Scholar 

  7. Rockwood K, Song X, MacKnight C, Bergman H, Hogan BD, McDowell I, Mitnitski A. A global clinical measure of fitness and frailty in elderly people. Can Med Assoc J. 2005;173(5):489–95.

    Article  Google Scholar 

  8. Barnett K, Mercer SW, Norbury M, Watt G, Wyke S, Guthrie B. Epidemiology of multimorbidity and implications for health care, research, and medical education: a cross-sectional study. Lancet. 2012;380(9836):37–43.

    Article  PubMed  Google Scholar 

  9. Rohrmann S. Epidemiology of frailty in older people. Frailty and cardiovascular diseases: research into an elderly population. 2020. p. 21–7.

    Book  Google Scholar 

  10. Hanlon P, Nicholl BI, Jani BD, Lee D, McQueenie R, Mair FS. Frailty and pre-frailty in middle-aged and older adults and its association with multimorbidity and mortality: a prospective analysis of 493 737 UK Biobank participants. Lancet Public Health. 2018;3(7):e323–32.

    Article  PubMed  PubMed Central  Google Scholar 

  11. Rodrigues LP, de Oliveira Rezende AT, Delpino FM, Mendonça CR, Noll M, Nunes BP, et al. Association between multimorbidity and hospitalization in older adults: systematic review and meta-analysis. Age Ageing. 2022;51(7):afac155.

    Article  PubMed  PubMed Central  Google Scholar 

  12. Veronese N, Siri G, Cella A, Daragjati J, Cruz-Jentoft AJ, Polidori MC, et al. Older women are frailer, but less often die than men: a prospective study of older hospitalized people. Maturitas. 2019;128:81–6.

    Article  PubMed  PubMed Central  Google Scholar 

  13. Kardas P, Ratajczyk-Pakalska E. Reasons for elderly patient hospitalization in departments of internal medicine in Lodz. Aging Clin Exp Res. 2003;15:25–31.

    Article  PubMed  Google Scholar 

  14. SBU. Kontinuitet i vården: En Systematisk översikt och utvärdering av medicinska, hälsoekonomiska och etiska aspekter. Continuity of care: a systematic review and assessment of medical, economic and ethical aspects. Stockholm: Statens beredning för medicinsk och social utvärdering (SBU); 2021.

    Google Scholar 

  15. WHO. World report on ageing and health. Geneva: World Health Organization; 2015.

  16. SCB. Befolkningsprognos för Sverige. Statistics Sweden. Projections of population growth of Sweden: Statistikmyndigheten SCB; 2022. Available from: https://www.scb.se/hitta-statistik/sverige-i-siffror/manniskorna-i-sverige/befolkningsprognos-for-sverige/ . Cited 2023 22 March.

  17. McCarthy D, Ryan J, Klein S. Models of care for high-need, high-cost patients: an evidence synthesis. New York: Commonwealth Fund; 2015.

    Google Scholar 

  18. Bodenheimer T, Berry-Millett R. Care management of patients with complex health care needs. Policy. 2009;1(6):1–16.

  19. Fried LP, Tangen CM, Walston J, Newman AB, Hirsch C, Gottdiener J, et al. Frailty in older adults: evidence for a phenotype. J Gerontol Ser A Biol Sci Med Sci. 2001;56(3):M146–57.

    Article  CAS  Google Scholar 

  20. Mitnitski AB, Mogilner AJ, Rockwood K. Accumulation of deficits as a proxy measure of aging. ScientificWorldJournal. 2001;1:323–36.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  21. Launer LJ, Wind AW, Deeg DJ. Nonresponse pattern and bias in a community-based cross-sectional study of cognitive functioning among the elderly. Am J Epidemiol. 1994;139(8):803–12.

    Article  CAS  PubMed  Google Scholar 

  22. López-Aguilà S, Contel JC, Farré J, Campuzano JL, Rajmil L. Predictive model for emergency hospital admission and 6-month readmission. Am J Manag Care. 2011;17(9):e348-357.

    PubMed  Google Scholar 

  23. Crane SJ, Tung EE, Hanson GJ, Cha S, Chaudhry R, Takahashi PY. Use of an electronic administrative database to identify older community dwelling adults at high-risk for hospitalization or emergency department visits: the elders risk assessment index. BMC Health Serv Res. 2010;10(1):1–7.

    Article  Google Scholar 

  24. Soong J, Poots A, Scott S, Donald K, Bell D. Developing and validating a risk prediction model for acute care based on frailty syndromes. BMJ Open. 2015;5(10):e008457.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  25. Coleman EA, Wagner EH, Grothaus LC, Hecht J, Savarino J, Buchner DM. Predicting hospitalization and functional decline in older health plan enrollees: are administrative data as accurate as self-report? J Am Geriatr Soc. 1998;46(4):419–25.

    Article  CAS  PubMed  Google Scholar 

  26. Tamang S, Milstein A, Sørensen HT, Pedersen L, Mackey L, Betterton J-R, et al. Predicting patient ‘cost blooms’ in Denmark: a longitudinal population-based study. BMJ Open. 2017;7(1):e011580.

    Article  PubMed  PubMed Central  Google Scholar 

  27. Klunder JH, Panneman SL, Wallace E, de Vries R, Joling KJ, Maarsingh OR, et al. Prediction models for the prediction of unplanned hospital admissions in community-dwelling older adults: a systematic review. PLoS One. 2022;17(9):e0275116.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  28. Kim H, Jung Y-i, Kh K. Prospectively identifying older adults at risk for potentially avoidable hospitalizations in Korea using population-based data. Int J Qual Health Care. 2019;31(8):620–6.

    PubMed  Google Scholar 

  29. Goodwin JS, Howrey B, Zhang DD, Kuo Y-F. Risk of continued institutionalization after hospitalization in older adults. Journals Gerontol Ser A: Biomedical Sci Med Sci. 2011;66(12):1321–7.

    Article  Google Scholar 

  30. Graf C. Functional decline in hospitalized older adults: it’s often a consequence of hospitalization, but it doesn’t have to be. AJN Am J Nurs. 2006;106(1):58–67.

    Article  PubMed  Google Scholar 

  31. Marcusson J, Nord M, Dong HJ, Lyth J. Clinically useful prediction of hospital admissions in an older population. BMC Geriatr. 2020;20(1):1–9.

    Article  Google Scholar 

  32. Nord M, Östgren CJ, Marcusson J, Johansson M. Staff experiences of a new tool for comprehensive geriatric assessment in primary care (PASTEL): a focus group study. Scand J Prim Health Care. 2020;38(2):132–45.

    Article  PubMed  PubMed Central  Google Scholar 

  33. Nord M, Lyth J, Alwin J, Marcusson J. Costs and effects of comprehensive geriatric assessment in primary care for older adults with high risk for hospitalisation. BMC Geriatr. 2021;21:1–9.

    Article  Google Scholar 

  34. Nord M, Lyth J, Marcusson J, Alwin J. Cost-effectiveness of Comprehensive Geriatric Assessment adapted to Primary Care. J Am Med Dir Assoc. 2022;23(12):2003–9.

    Article  PubMed  Google Scholar 

  35. WHO. International statistical classification of diseases and related health problems, 10th revision, Fifth edition. Geneva: World Health Organization; 2016.

  36. Gray RJ. A class of K-sample tests for comparing the cumulative incidence of a competing risk. Annals Stat. 1988;16:1141–54.

  37. Fine JP, Gray RJ. A proportional hazards model for the subdistribution of a competing risk. J Am Stat Assoc. 1999;94(446):496–509.

    Article  MathSciNet  Google Scholar 

  38. WHO. International statistical classification of diseases and related health problems 10th revision volume 2 instruction manual fifth edition. Geneva: World Health Organization; 2016.

  39. Kojima G, Iliffe S, Walters K. Frailty index as a predictor of mortality: a systematic review and meta-analysis. Age Ageing. 2018;47(2):193–200.

    Article  PubMed  Google Scholar 

  40. Clegg A, Bates C, Young J, Ryan R, Nichols L, Ann Teale E, et al. Development and validation of an electronic frailty index using routine primary care electronic health record data. Age Ageing. 2016;45(3):353–60.

    Article  PubMed  PubMed Central  Google Scholar 

  41. Juma S, Taabazuing M-M, Montero-Odasso M. Clinical frailty scale in an acute medicine unit: a simple tool that predicts length of stay. Can Geriatr J. 2016;19(2):34.

    Article  PubMed  PubMed Central  Google Scholar 

  42. Haley MN, Wells YD, Holland AE. Relationship between frailty and discharge outcomes in subacute care. Aust Health Rev. 2013;38(1):25–9.

    Article  Google Scholar 

  43. Wang HT, Fafard J, Ahern S, Vendittoli P-A, Hebert P. Frailty as a predictor of hospital length of stay after elective total joint replacements in elderly patients. BMC Musculoskelet Disord. 2018;19(1):1–6.

    Article  Google Scholar 

  44. Khandelwal D, Goel A, Kumar U, Gulati V, Narang R, Dey AB. Frailty is associated with longer hospital stay and increased mortality in hospitalized older patients. J Nutr Health Aging. 2012;16(8):732–5.

    Article  CAS  PubMed  Google Scholar 

  45. Evans SJ, Sayers M, Mitnitski A, Rockwood K. The risk of adverse outcomes in hospitalized older patients in relation to a frailty index based on a comprehensive geriatric assessment. Age Ageing. 2014;43(1):127–32.

    Article  PubMed  Google Scholar 

  46. Passey ME, Longman JM, Johnston JJ, Jorm L, Ewald D, Morgan GG, et al. Diagnosing potentially preventable hospitalisations (DaPPHne): protocol for a mixed-methods data-linkage study. BMJ Open. 2015;5(11):e009879.

    Article  PubMed  PubMed Central  Google Scholar 

  47. WHO. Assessing health services delivery performance with hospitalizations for ambulatory care sensitive conditions. World Health Organization. Regional Office for Europe; 2016. https://iris.who.int/bitstream/handle/10665/348972/WHO-EURO-2016-4172-43931-61907-eng.pdf?sequence=3&isAllowed=y. Cited 2023 May 03.

  48. Socialstyrelsen. Utveckling av indikatorerna undvikbar slutenvård och oplanerade återinskrivningar. National Board of Health and Welfare. Development of the indicators avoidable inpatient care and unplanned readmissions. Stockholm: Socialstyrelsen; 2014. https://www.socialstyrelsen.se/globalassets/sharepointsharepoint-dokument/artikelkatalog/ovrigt/2014-2-12.pdf. Cited 2023 March 22.

  49. Data Twb. Hopsital beds (per 1,000 people) the world bank. Available from: https://data.worldbank.org/indicator/SH.MED.BEDSZS?end=2019&name_desc=false&start=1960. Cited 2023 March 22.

  50. Data O. Length of hospital stay: the organisation for economic co-operation and development. Available from: https://data.oecd.org/healthcare/length-of-hospital-stay.htm . Cited 2023 March 22.

Download references

Acknowledgements

We thank Jan Marcusson for support when conducting this study.

Funding

Open access funding provided by Linköping University. This study was performed by the authors while employed by Linköping University and/or Region Östergötland. The work was supported by Region Östergötland and Linköping University from the strategic research fund for ‘Health Care and Welfare’ [Grant number 2016186–14], Swedish Research Council (ALF) [Grant number RÖ-974820], and the Research Council of Southeastern Sweden [Grant number FORSS-969444]. The funders had no input or influence on the study.

Author information

Authors and Affiliations

  1. Primary Health Care Centre Finspång, Finspång, Sweden

    Moa Lundgren

  2. Department of Health, Medicine and Caring Sciences, Linköping University, Linköping, Sweden

    Moa Lundgren, Anna Segernäs, Magnus Nord, Jenny Alwin & Johan Lyth

  3. Primary Health Care Centre Ekholmen, Linköping, Sweden

    Anna Segernäs

  4. Primary Health Care Centre Valla, Linköping, Sweden

    Magnus Nord

Authors
  1. Moa Lundgren
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Anna Segernäs
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Magnus Nord
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Jenny Alwin
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Johan Lyth
    View author publications

    You can also search for this author in PubMed Google Scholar

Contributions

AS, JL, JA and ML designed the study. JL performed the statistical analysis. AS, JL, JA and ML interpreted the data and participated in the framework construction of the manuscript. ML wrote the first draft of the manuscript. AS, JL, JA and MN read, improved and approved the final version of the manuscript.

Corresponding author

Correspondence to Moa Lundgren.

Ethics declarations

Ethics approval and consent to participate

The study was conducted in accordance with the principles of the Declaration of Helsinki. The study was approved by The regional ethical review board in Linköping (Dnr 2016/347–31). In the ethics application, the study design without informed consents was described which the regional ethical review board in Linköping (Dnr 2016/347–31) approved. This is in accordance with item 32 in the declaration.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Lundgren, M., Segernäs, A., Nord, M. et al. Reasons for hospitalisation and cumulative mortality in people, 75 years or older, at high risk of hospital admission: a prospective study. BMC Geriatr 24, 176 (2024). https://doi.org/10.1186/s12877-024-04771-2

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1186/s12877-024-04771-2

Keywords

BMC Geriatrics

ISSN: 1471-2318

Contact us