Reasons for home care clients’ unplanned Hospital admissions and their associations with patient characteristics

Reasons for Home Care Clients’ Unplanned Hospital Admissions and Their Associations with Patient Characteristics Jukka K. Rönneikkö, MD;a,* Esa R. Jämsen, MD, PhD;b,c,d Matti Mäkelä, MD, PhD;e Harriet Finne-Soveri, MD, PhD;f Jaakko N. Valvanne MD, PhDb,c,g aYlöjärvi Health Care Centre, Ylöjärvi, Finland; bFaculty of Medicine and Life Sciences, University of Tampere, Tampere, Finland; Gerontology cResearch Center (GEREC), Universities of Jyväskylä and Tampere, Finland; Hatanpää hospital, dService line of general practice and geriatrics, Tampere, Finland; eNational Institute for Health and Welfare, Helsinki, Finland; fElderly care services, City of Helsinki, Finland; gPromotion of Senior Citizens’ Welfare, City of Tampere, Finland; Tampere University Hospital, Tampere, Finland *Address correspondence to Jukka K. Rönneikkö, Ylöjärvi Health Centre, Mikkolantie 10, 33470 Ylöjärvi Finland. E-mail [email protected] Abbreviated title: Hospitalizations among New Home Care Clients © 2018 Elsevier. This manuscript version is made available under the CC-BY-NC-ND 4.0 license http://creativecommons.org/licenses/by-nc-nd/4.0/ This is the post print version of the article published in Archives of gerontology and geriatrics 2018, 78, 114-126. The final publication is available at https://doi.org/10.1016/j.archger.2018.06.008 2 BACKGROUND: Unplanned hospitalizations and emergency room visits occur frequently among home care clients The aim of this study was to identify typical discharge diagnoses and their associations with patient characteristics among a total of 6,812 Finnish home care clients aged ≥63 years who were hospitalized within one year of their first home care assessment. METHODS: A register-based study based on Resident Assessment Instrument-Home Care (RAI-HC) assessments and nationwide hospital discharge records. The RAI-HC assessments were linked to the hospital discharge records of the participants’ first unplanned hospitalization. Univariate and multivariable regression analyses were used to evaluate the association of RAI-HC determinants with discharge diagnoses. RESULTS: The most common reason for the first hospitalization was an infectious disease (21%; n=1,446). When hospitalizations were classified according to the main diagnosis, chronic skin ulcers, functional impairment and daily urinary incontinence were associated with hospitalization due to infectious diseases; impaired cognitive capacity, Alzheimer’s disease or other dementia and polypharmacy (protective effect) were associated with hospitalizations due to dementia; age of ≥90 years, congestive heart failure, coronary artery disease and using ≥10 drugs with hospitalizations due to heart diseases; and moderate or strong pain with hospitalization due to musculoskeletal disorders. Previous falls, female sex and an earlier hip fracture were associated with injury-related hospitalizations. Feelings of loneliness increased the odds of hospitalization due to geriatric symptoms without a specific diagnosis. CONCLUSION: Patient characteristics and geriatric syndromes identified using RAI-HC predict the reasons for future hospitalizations among new home care clients. Key words: hospitalization, RAI assessment, home care, diagnosis 3 1. INTRODUCTION The aim of home care services is to help people with functional limitations to live in their own homes. Home care is considered a possible strategy for reducing hospital use among older people (1), and in fact, there are studies showing that home care prevents hospitalizations among selected older people (2, 3). Nevertheless, unplanned hospitalizations and emergency room visits occur frequently among home care clients, and they are often associated with adverse outcomes (4-6). Although hospitalizations are often due to acute exacerbations of chronic diseases (7), an earlier study among new home care clients indicated that many of the risk factors predicting unplanned hospitalization represent major geriatric challenges (6). Targeting the identified risk factors for hospitalization (6-12) could provide a means to prevent future hospitalizations. A better understanding of how patient characteristics are linked to different reasons for hospitalization would help to identify potentially modifiable conditions and thereby to reduce hospital admissions. The aim of the present study, which is based on nationwide register data and Resident Assessment Instrument for Home Care (RAI-HC) assessments, is therefore to identify conditions that could be targeted in the care planning of home care clients to prevent hospital admissions. This study extends an earlier analysis (6), describes the most common discharge diagnoses, and analyzes how patient characteristics are associated with the reasons for hospitalization. 4 2. METHODS The Resident Assessment Instrument for Home Care (RAI-HC) is a comprehensive assessment system developed by InterRAI, a multinational research network dedicated to cross-national comparisons of health and health care for elderly people and people with disabilities. The collected data contains core assessment items necessary for a comprehensive assessment, such as function, health, social support, and service use (13), and its reliability and validity have been tested in international studies (2, 13, 14). This study was based on the first RAI-HC assessments made for new home care clients (i.e. persons with no previous use of home care), aged ≥63 years, in Finland in 2009–2011 (n=15,700) and hospital discharge records of those clients who were hospitalized at least once within one year of the first RAI-HC assessment (n=6,812). The nurses responsible for each client perform the assessments, and they have been trained in the use of RAI-HC. According to national guidelines, home care clients are assessed upon initiation of services and thereafter at least twice a year. The RAI data were collected from the national database maintained by the National Institute for Health and Welfare (THL), which includes assessments from both rural (30%) and urban (70%) settings. Of the RAI-HC scales, those measuring activities of daily living performance (ADLh) (15), cognitive performance (CPS) (16), depression (DRS) (17), pain, and health stability (CHESS) (18) were used in this study. The variables used in the analyses are listed in Table 1. The information about hospitalizations and discharge diagnoses (according to International Classification of Diagnoses, 10th revision) were collected from the Finnish Hospital Discharge Register (FHDR) and were linked to the RAI-HC data using each citizen’s unique identification number. The FHDR contains the data of all discharged patients from inpatient care in health centers and hospitals, including both public and private institutions. The coverage of the register exceeds 95%, and the completeness and accuracy of the registered items varies from satisfactory to very good (19). 5 For this study, only the first hospitalizations occurring within one year of the first RAI-HC assessment were included. Scheduled hospitalizations (e.g. elective surgery) were excluded, because the aim was to analyze unplanned hospitalizations. The hospitalizations were divided into nine diagnosis groups according to their primary discharge diagnoses (the first registered diagnosis representing the main cause of hospitalization according to the treating physician): infectious diseases; dementia; cardiovascular, cerebrovascular, and musculoskeletal diseases; other specific diseases; geriatric symptoms (e.g. malaise, dizziness, syncope, malnutrition); injuries; and other reasons (Appendix Table A1). The diagnosis groups were determined according to the previous studies concerning hospital care among old people (20, 21) and to the authors’ clinical experience. Final classification was reached by consensus between three experienced geriatricians. Finally, in order to clarify how often geriatric syndromes are identified and registered in discharge records, we searched both primary and secondary diagnoses for records about malnutrition and delirium. 2.1. Ethics The use of the RAI database for this study, including linkage to the FHDR, was approved by the THL ethics committee (THL/134/6.02.00/2012; Feb 16, 2012). 2.2. Statistical analyses A two-stage analysis was used to identify associations between the status of new home care clients and the reasons for their later hospitalization. First, a univariate analysis was used to explore associations between RAI-HC determinants and diagnosis groups. In the second stage, multivariable models were constructed in order to identify factors with independent association with the causes of hospitalizations. All variables with p ≤0.01 in the univariate analysis were included in these multivariable models. This level of significance was selected because of the high number of separate comparisons. Although the analyses were performed separately for all nine diagnosis groups, all variables found to have statistically significant 6 association with hospitalization in the univariate analyses (independent of the reason for hospitalization) were included in all multivariable models, because these all were considered clinically important and potentially confounding factors in all diagnosis groups, despite some not being significant for each of the diagnosis groups individually. The results are presented as odds ratios (OR) with 95% confidence intervals (CI). The odds ratios indicate how much higher (or lower) the possibility of the analyzed diagnosis group being the reason for hospitalization (e.g. infectious disease or injury) is in patients with a specific characteristic (e.g. age, diagnosis of dementia, ADL performance) than in patients without that characteristic. All the statistical analyses were performed using SAS 9.3 (SAS Institute INC., Cary, NC, USA). 7 3. RESULTS Of the 15,700 new home care clients, 6,812 (43%) were hospitalized. Some 48 patients (0.7%) were excluded because of missing discharge diagnoses. The characteristics of the remaining 6,764 patients are described in Table 1. The most common reasons for hospitalization were infectious diseases (21.2%; n=1,446). In 34% of hospitalizations due to infectious diseases, the specific diagnosis was urinary tract infection, and it was the most common individual diagnosis (7.2%; n=490) (Table 2). Dementia disorders were the reason for hospitalization in 449 patients (6.6%). Other geriatric syndromes were recorded very infrequently: delirium was recorded as the main diagnosis in 11 patients and as the secondary diagnosis in 8 patients. Malnutrition as a main or secondary diagnosis was recorded not even once. 3.1. Univariate analysis Most of the analyzed variables were associated with some of the diagnosis groups in the univariate analysis (Table 3). Only living alone, housing defects, vision, dementia associated with Parkinson's disease, a psychiatric diagnosis, and psychotropic medication had no significant association (p ≤0.01) with any diagnosis group, and they were therefore excluded from the multivariable models. 3.2. Multivariable analysis The results of the multivariable analysis are shown in Table 4. When hospitalizations were classified according to the main diagnosis, chronic skin ulcers (OR 1.35, 95% CI 1.11–1.65), functional impairment (ADLh 1–2 1.23, 1.04–1.47; ADLh 3–4 1.40, 1.12–1.75; ADLh 5–6 1.65, 1.11–2.50), and daily urinary incontinence (1.24, 1.05–1.45) were associated with hospitalizations due to infectious diseases. Impaired cognitive capacity (CPS 1–2: 4.10, 2.72– 6.18; CPS 3–4: 4.70, 2.84–7.78; CPS 5–6: 5.66, 3.27–9.8), Alzheimer’s disease (1.77, 1.31– 2.38), and other dementia disorders (2.63, 1.91–3.64) increased the likelihood that the reason 8 for hospitalization was dementia, while polypharmacy was associated with a decreased probability of dementia-related hospitalization (5–9 drugs: 0.69, 0.52–0.91; ≥10 drugs: 0.47, 0.33–0.66). A history of heart diseases (congestive heart failure (2.40, 2.05–2.82) and coronary artery disease (1.71, 1.46–2.00)) was associated with hospitalizations due to heart diseases. A similar effect was found for an age of ≥90 years (2.03, 1.53–2.70) and using ≥10 drugs (1.37, 1.01–1.86). Moderate or strong pain (1.76, 1.35–2.29) was more common among patients with hospitalization due musculoskeletal disorders. Previous falls (1.43, 1.21–1.68), female sex (1.33, 1.11–1.60), and an earlier hip fracture (1.61, 1.11–2.32) were associated with injury-related hospitalizations. Feeling lonely increased the odds of being hospitalized due to geriatric symptoms without a specific diagnosis (1.35, 1.08–1.69). Despite statistically significant analyses in the multivariate models, only a minority of hospitalizations among patients with a certain characteristic were due to the cause of hospitalization that the characteristic was associated with. One third of hospitalizations among patients with ADL score of 5-6 and approximately one-fourth (24-29%) of hospitalizations among patients with ADL 1-4, with wounds or urinary incontinence, were due to infections.. Dementia accounted for one-fifth (17-27%) of hospitalizations among patients with cognitive decline, Alzheimer’s disease or other dementia. The proportion was similar for associations related to hospitalizations due to heart diseases, whereas moderate pain explained only 6% of hospitalizations for musculoskeletal disorders, and female sex, previous falls and history of hip fracture about 14-19% of hospitalizations due to injuries. 9 4. DISCUSSION This study indicates that already at the beginning of home care services, certain patient characteristics are related to specific discharge diagnoses if the new home care client becomes hospitalized within one year of the first RAI-HC assessment. These associations differ partly from the risk factors for hospitalization identified in early studies (6-12), because only patients who were hospitalized were included in this study. Infectious diseases were the most common reason for hospitalization, and urinary tract infections the most common individual diagnosis. As expected, some common symptoms and geriatric syndromes, for example delirium and malnutrition, were rarely documented probably because of inadequate identification and registration of these conditions in clinical practice (22). Despite the observed significant associations, it however turned out that among all patient subgroups the reasons for hospitalizations varied markedly, and at maximum one-third of hospitalizations were due to a certain disease-group (i.e. two-thirds were due to conditions in the other eight diagnosis groups). In an earlier study, previous hospitalizations represented the strongest independent risk factor for unplanned hospitalization after the initiation of home care services (6). Interestingly, previous hospitalizations were not associated with dementia-related hospitalizations or injuries. Instead, they were associated with an increased probability of being hospitalized due to cardiovascular diseases or other chronic conditions. In the univariate analysis, CHESS ≥1 as a marker of unstable health status also increased the likelihood of hospitalization for specific diagnoses, including malignant neoplasms and diseases of the nervous, circulatory, metabolic, respiratory, or digestive system. In new home care clients with previous hospitalizations, special attention should be paid to good care and the monitoring of underlying chronic conditions. Furthermore, not only diseases but also their symptoms require 16 ACKNOWLEDGMENTS The authors would like to thank interRAI for the opportunity to use the RAI-HC assessment tool and the Finnish home care personnel for conducting the RAI-HC assessments. Conflict of interest statement Financial/Personal Conflicts Rönneikkö Jämsen Mäkelä FinneSoveri Valvanne Yes No Yes No Yes No Yes No Yes No Employment or Affiliation X X X X X Grants/Funds X X X X X Honoraria X X X X X Speaker Forum X X X X X Consultant X X X X X Stocks X X X X X Royalties X X X X X Expert Testimony X X X X X Board Member X X X X X Patents X X X X X Personal Relationship X X X X X Jämsen: Payment for lectures (unrelated to the present study) by the Orion Corporation (Espoo, Finland) and Novartis Finland (Espoo, Finland). Funding: This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. 17 References 1. Landi F, Onder G, Russo A et al. A new model of integrated home care for the elderly: impact on hospital use. J Clin Epidemiol 2001;54:968-970. 2. Bernabei R, Landi F, Gambassi G et al. Randomised trial of impact of model of integrated care and case management for older people living in the community. BMJ 1998;316:1348-1351. 3. Vasquez MS. Preventing rehospitalization through effective home health nursing care. Home Healthc Nurse 2008;26:75-81. 4. Doran DM, Hirdes J, Poss J et al. Identification of safety outcomes for Canadian home care clients: evidence from the resident assessment instrument--home care reporting system concerning emergency room visits. Healthc Q 2009;12:40-48. 5. Dorr DA, Jones SS, Burns L et al. Use of health-related, quality-of-life metrics to predict mortality and hospitalizations in community-dwelling seniors. J Am Geriatr Soc 2006;54:667-673. 6. Ronneikko JK, Makela M, Jamsen ER et al. Predictors for Unplanned Hospitalization of New Home Care Clients. J Am Geriatr Soc 2017;65:407-414. 7. Landi F, Onder G, Cesari M et al. Comorbidity and social factors predicted hospitalization in frail elderly patients. J Clin Epidemiol 2004;57:832-836. 8. Rosati RJ, Huang L, Navaie-Waliser M et al. Risk factors for repeated hospitalizations among home healthcare recipients. Journal for Healthcare Quality 2003;25:4-10. 9. Shugarman LR, Buttar A, Fries BE et al. Caregiver attitudes and hospitalization risk in michigan residents receiving homeand community-based care. J Am Geriatr Soc 2002;50:1079-1085. 18 10. Paddock K, Hirdes JP. Acute health care service use among elderly home care clients. Home Health Care Serv Q 2003;22:75-85. 11. Fortinsky RH, Madigan EA, Sheehan TJ et al. Risk factors for hospitalization among Medicare home care patients. West J Nurs Res 2006;28:902-917. 12. Morris JN, Howard EP, Steel K et al. Predicting risk of hospital and emergency department use for home care elderly persons through a secondary analysis of cross-national data. BMC Health Services Research 2014;14:519. 13. Morris JN, Fries BE, Steel K et al. Comprehensive clinical assessment in community setting: applicability of the MDS-HC. J Am Geriatr Soc 1997;45:1017-1024. 14. Landi F, Tua E, Onder G et al. Minimum data set for home care: a valid instrument to assess frail older people living in the community. Med Care 2000;38:1184-1190. 15. Morris JN, Fries BE, Morris SA. Scaling ADLs within the MDS. Journals of Gerontology Series ABiological Sciences & Medical Sciences 1999;54:M546-53. 16. Morris JN, Fries BE, Mehr DR et al. MDS Cognitive Performance Scale. J Gerontol 1994;49:M17482. 17. Burrows AB, Morris JN, Simon SE et al. Development of a minimum data set-based depression rating scale for use in nursing homes. Age & Ageing 2000;29:165-172. 18. Hirdes J, Frijters D, Teare G. The MDS-CHESS scale: A new measure to predict mortality in institutionalized older people. J Am Geriatr Soc 2003;51:96-100. 19. Sund R. Quality of the Finnish Hospital Discharge Register: a systematic review. Scand J Public Health 2012;40:505-515. 19 20. Walsh B, Roberts H, Hopkinson J. Emergency hospital admissions for ill-defined conditions amongst older people: a review of the literature.. International Journal of Older People Nursing 2007;2:270-277. 21. Natalwala A, Potluri R, Uppal H et al. Reasons for hospital admissions in dementia patients in Birmingham, UK, during 2002-2007. Dementia & Geriatric Cognitive Disorders 2008;26:499-505. 22. Buurman BM, Hoogerduijn JG, de Haan RJ et al. Geriatric conditions in acutely hospitalized older patients: prevalence and one-year survival and functional decline. PLoS ONE [Electronic Resource] 2011;6:e26951. 23. Colt HG, Shapiro AP. Drug-induced illness as a cause for admission to a community hospital. J Am Geriatr Soc 1989;37:323-326. 24. Flaherty JH, Perry HM3, Lynchard GS et al. Polypharmacy and hospitalization among older home care patients. Journals of Gerontology Series A-Biological Sciences & Medical Sciences 2000;55:M554-9. 25. Col N, Fanale JE, Kronholm P. The role of medication noncompliance and adverse drug reactions in hospitalizations of the elderly. Arch Intern Med 1990;150:841-845. 26. Helgadottir B, Laflamme L, Monarrez-Espino J et al. Medication and fall injury in the elderly population; do individual demographics, health status and lifestyle matter?. BMC Geriatrics 2014;14:92. 27. Fried LP, Tangen CM, Walston J et al. Frailty in older adults: evidence for a phenotype. Journals of Gerontology Series A-Biological Sciences & Medical Sciences 2001;56:M146-56. 28. Boyd CM, Xue Q, Simpson CF et al. Frailty, hospitalization, and progression of disability in a cohort of disabled older women. Am J Med 2005;118:1225-1231. 20 29. Woodford HJ, George J. Diagnosis and management of urinary tract infection in hospitalized older people. J Am Geriatr Soc 2009;57:107-114. 30. Ojanen R, Huttunen R, Syrjanen J et al. Urinary tract infections in acutely hospitalized geriatric patients. Finnish Medical Journal 2016;71:2635-2643. 31. Schulz L, Hoffman RJ, Pothof J et al. Top Ten Myths Regarding the Diagnosis and Treatment of Urinary Tract Infections. J Emerg Med 2016;51:25-30. 32. Inouye SK. Delirium in hospitalized older patients: recognition and risk factors. Journal of Geriatric Psychiatry & Neurology 1998;11:118-125. 33. Miu DKY, Chan CW, Kok C. Delirium among elderly patients admitted to a post-acute care facility and 3-months outcome. Geriatrics & Gerontology International 2016;16:586-592. 34. Siddiqi N, House AO, Holmes JD. Occurrence and outcome of delirium in medical in-patients: a systematic literature review. Age & Ageing 2006;35:350-364. 35. Meuleners LB, Fraser ML, Bulsara MK et al. Risk factors for recurrent injurious falls that require hospitalization for older adults with dementia: a population based study. BMC Neurology 2016;16:188. 36. Close J, Ellis M, Hooper R et al. Prevention of falls in the elderly trial (PROFET): a randomised controlled trial. Lancet 1999;353:93-97. 37. Davison J, Bond J, Dawson P et al. Patients with recurrent falls attending Accident & Emergency benefit from multifactorial intervention--a randomised controlled trial. Age Ageing 2005;34:162-168. 38. Pitkala KH, Routasalo P, Kautiainen H et al. Effects of psychosocial group rehabilitation on health, use of health care services, and mortality of older persons suffering from loneliness: a randomized, 21 controlled trial. Journals of Gerontology Series A-Biological Sciences & Medical Sciences 2009;64:792800. 39. Pavlou M, Ambler G, Seaman S et al. Review and evaluation of penalised regression methods for risk prediction in low-dimensional data with few events. Stat Med 2016;35:1159-1177. 40. Peduzzi P, Concato J, Kemper E et al. A simulation study of the number of events per variable in logistic regression analysis. J Clin Epidemiol 1996;49:1373-1379. 22 Table 1. Characteristics of those hospitalized ALL N % 6812 Demographic Age 63-74v 1160 17,0 75-79v 1137 16,7 80-84v 1935 28,4 85-89v 1675 24,6 90+v 905 13,3 Gender Female 4610 67,7 Male 2202 32,3 Social situation Living alone 4624 67,9 Caregiver stressed 611 9,0 Housing defects 1723 25,3 Use of services Reason for home care: client has been discharged from hospital 2384 35,0 Hospitalization during one year before assessment 5135 75,4 Acute outpatient care in 90 days before assessment 2537 37,2 Function ADLH (0 - 6) 0 4920 72,2 1-2 1079 15,8 3-4 669 9,8 5-6 144 2,1 Client believes he/she is capable of improving performance in physical function 1117 16,4 CPS 0 2599 38,2 1-2 3317 48,7 3-4 514 7,5 5-6 382 5,6 Clinical symptoms Urinary incontinence daily 1319 19,4 Fecal incontinency 537 7,9 Chronic skin ulcers 586 8,6 23 Mouth problems 5637 82,8 Vision good enough 5076 74,5 moderately impaired 1557 22,9 severely impaired 179 2,6 Falls during 90 days before assessment 2171 31,9 DRS 0-2 5669 83,2 3-14 1143 16,8 Feeling lonely 1554 22,8 Poor self-rated health 2307 33,9 BMI <18.5 404 5,9 18.5-23.9 2251 33,0 24-29.9 2279 33,5 ≥ 30 1076 15,8 PAIN 0-1 4378 64,3 2-3 2434 35,7 Diagnoses Congestive heart failure 1502 22,0 Coronary artery disease 1808 26,5 Alzheimer’s disease 1299 19,1 Other dementia 1932 28,4 Old stroke 1932 28,4 Parkinson’s disease 261 3,8 Parkinson’s disease and dementia 66 1,0 Musculoskeletal disorders 1711 25,1 Old hip fracture 242 3,6 Old other fracture 262 3,8 Cancer 644 9,5 Renal insufficiency 409 6,0 Psychiatric diagnosis 852 12,5 Chronic obstructive pulmonary disease 832 12,2 Diabetes 1632 24,0 Medication Number of drugs * 0-4 734 10,8 5-9 2864 42,0 10 or more 3214 47,2 Psychotropic medication 3638 53,4 Influenza vaccination 3687 54,1 Health stability 24 CHESS 0 3026 44,4 1 1857 27,3 2-5 1929 28,3 * including prescription and non-prescription medications ADLh =The Activities of Daily Living Hierarchy CPS=The Cognitive Performance Scale DRS= The Depression Rating Scale PAIN=The Pain Scale CHESS= The Changes in Health, End-Stage Disease, Signs, and Symptoms Scale 25 Table 2. The diagnosis groups and ten most common discharge diagnoses of those hospitalized N % The diagnosis groups Infectious diseases 1446 21.2 Dementia diseases 449 6.6 Cardiovascular diseases 905 13.3 Cerebrovascular diseases 311 4.6 Musculoskeletal diseases 280 4.1 Injuries 778 11.4 Other specific diseases 964 14.1 Geriatric symptoms 487 7.1 Other diseases and symptoms 1144 16.9 Ten most common discharge diagnoses and their ICD10 codes Urinary tract infection (N10, N30, N39.0) 490 7,2 Pneumonia (J12-J18) 435 6,4 Congestive heart failure (I50) 368 5,4 Alzheimer's disease (G30) 343 5,0 Hip fracture (S72) 242 3,6 Malignant neoplasms (C00-C26, C30-C41, C43-C58, C60-C97) 237 3,5 Malaise and fatigue (R53) 195 2,9 Atrial fibrillation and flutter (I48) 180 2,6 Cerebral infarction (I63) 171 2,5 Acute myocardial infarction (I21) 158 2,3 32 Additional Supporting Information may be found in the online version of this article: https://doi.org/10.1016/j.archger.2018.06.008