Heterogeneous length-of-stay modeling of post-acute care residents in the nursing home with competing discharge dispositions

doi:10.1007/s42524-022-0203-7

Front. Eng

2022, Vol. 9

Issue (4) : 577-591 https://doi.org/10.1007/s42524-022-0203-7

RESEARCH ARTICLE

Heterogeneous length-of-stay modeling of post-acute care residents in the nursing home with competing discharge dispositions

Nazmus SAKIB¹, Xuxue SUN², Nan KONG³, Chris MASTERSON⁴, Hongdao MENG⁵, Kelly SMITH⁵, Mingyang LI¹(

)

¹. Department of Industrial and Management Systems Engineering, University of South Florida, Tampa, FL 33620, USA
². College of Media Engineering, Communication University of Zhejiang, Hangzhou 310019, China
³. Weldon School of Biomedical Engineering, Purdue University, West Lafayette, IN 47907, USA
⁴. Greystone Health Network, Tampa, FL 33610, USA
⁵. School of Aging Studies, University of South Florida, Tampa, FL 33620, USA

Download: PDF(2141 KB) HTML
Export: BibTeX | EndNote | Reference Manager | ProCite | RefWorks

Abstract

Post-acute care (PAC) residents in nursing homes (NHs) are recently hospitalized patients with medically complex diagnoses, ranging from severe orthopedic injuries to cardiovascular diseases. A major role of NHs is to maximize restoration of PAC residents during their NH stays with desirable discharge outcomes, such as higher community discharge likelihood and lower re/hospitalization risk. Accurate prediction of the PAC residents’ length-of-stay (LOS) with multiple discharge dispositions (e.g., community discharge and re/hospitalization) will allow NH management groups to stratify NH residents based on their individualized risk in realizing personalized and resident-centered NH care delivery. Due to the highly heterogeneous health conditions of PAC residents and their multiple types of correlated discharge dispositions, developing an accurate prediction model becomes challenging. Existing predictive analytics methods, such as distribution-/regression-based methods and machine learning methods, either fail to incorporate varied individual characteristics comprehensively or ignore multiple discharge dispositions. In this work, a data-driven predictive analytics approach is considered to jointly predict the individualized re/hospitalization risk and community discharge likelihood over time in the presence of varied residents’ characteristics. A sampling algorithm is further developed to generate accurate predictive samples for a heterogeneous population of PAC residents in an NH and facilitate facility-level performance evaluation. A real case study using large-scale NH data is provided to demonstrate the superior prediction performance of the proposed work at individual and facility levels through comprehensive comparison with a large number of existing prediction methods as benchmarks. The developed analytics tools will allow NH management groups to identify the most at-risk residents by providing them with more proactive and focused care to improve resident outcomes.

Keywords nursing home predictive analytics individualized prediction competing risks health outcomes

Corresponding Author(s): Mingyang LI

Just Accepted Date: 16 September 2022 Online First Date: 07 November 2022 Issue Date: 08 December 2022

Cite this article:

Nazmus SAKIB,Xuxue SUN,Nan KONG, et al. Heterogeneous length-of-stay modeling of post-acute care residents in the nursing home with competing discharge dispositions[J]. Front. Eng, 2022, 9(4): 577-591.

URL:

https://academic.hep.com.cn/fem/EN/10.1007/s42524-022-0203-7
https://academic.hep.com.cn/fem/EN/Y2022/V9/I4/577

Proposed Sampling Algorithm
Step 1:	Compute $? i (t \| x i) = ∑ ? s ∈ {C, H} d^s b (t) exp ? (β^s T x i)$ , where $? i (t \| x i)$ is the instantaneous probability of resident i with $x i$ being discharged at time t;
Step 2:	Compute $S i (l) = exp ? [? ∑ p = 0 l ? i (t (p) \| x i)]$ , $l = 0, . . ., N, N + 1$ , where $S i (l)$ is the probability of resident i still residing in NH at time $t (l)$ ; $t (0) < t (1) < . . . < t (l) < . . . < t (N) < t (N + 1)$ and $t (0) = 0$ , $t (N + 1) = + ∞$ ; ${t (l)} l = 1 n$ are ordered distinct historical LOS observations;
Step 3:	Randomly generate $μ = U n i f (0, 1)$ ;
Step 4:	Compute $l 1 = max {l : μ ? S i (l)}$ , $l 2 = min {l : min S i (l) ? μ}$ , and get simulated LOS, $T i$ as $T i = S i (l) ? μ S i (l 1) ? S i (l 2) ? (t (l 2) ? t (l 1)) + t (l 1)$ ;
Step 5:	Determine disposition state s by drawing for the categorical distribution as $w s = d^s b (t) exp ? (β^s T x i) ? i (t \| x i)$ , $s ∈ {C, H}$ , i.e., $s ～ C a t e g o r i c a l (w)$ , where $w = [w C, w H] T$

Tab.1 Descriptive summary statistics of the selected resident cohort

Covariate short name	MDS 3.0 covariate description
Age	Age at admission (years)
ADL score	Activities of daily living score at admission
Cognitive score	Brief interview for mental status (BIMS) summary score at admission
Mood score	Resident mood interview patient health questionnaire (PHQ)-9 total severity score at admission
	Active diagnose indicator of resident at admission
Cancer	Cancer (with or without metastasis)
Anemia	Anemia (e.g., aplastic, iron deficiency, pernicious, and sickle cell)
Atrial fibrillation	Atrial fibrillation or other dysrhythmias (e.g., bradycardias and tachycardias)
Coronary artery disease	Coronary artery disease (CAD) (e.g., angina, myocardial infarction, and atherosclerotic heart disease (ASHD))
DVT, PE, PTE	Deep venous thrombosis (DVT), Pulmonary embolus (PE), or Pulmonary thrombo-embolism (PTE)
Heart failure	Heart failure (e.g., congestive heart failure (CHF) and pulmonary edema)
Hypertension	Hypertension
GERD or Ulcer	Gastroesophageal reflux disease (GERD) or Ulcer (e.g., esophageal, gastric, and peptic ulcers)
UC, CD, IBD	Ulcerative colitis (UC), Crohn’s disease (CD), or Inflammatory bowel disease (IBD)
BPH	Benign prostatic hyperplasia (BPH)
Renal disease	Renal insufficiency, Renal failure, or End-stage renal disease (ESRD)
Neurogenic bladder	Neurogenic bladder
Obstructive uropathy	Obstructive uropathy
MDRO	Multidrug-resistant organism (MDRO)
Pneumonia	Pneumonia
Septicemia	Septicemia
Diabetes mellitus	Diabetes mellitus (DM) (e.g., diabetic retinopathy, nephropathy, and neuropathy)
Hyponatremia	Hyponatremia
Hyperlipidemia	Hyperlipidemia (e.g., hypercholesterolemia)
Arthritis	Arthritis (e.g., degenerative joint disease (DJD), osteoarthritis, and rheumatoid arthritis (RA))
Hip fracture	Hip fracture (e.g., sub-capital fractures, and fractures of the trochanter and femoral neck)
Other fracture	Other fracture
Alzheimer’s disease	Alzheimer’s disease
Non-Alzheimer’s dementia	Non-Alzheimer’s dementia (e.g., Lewy body dementia, vascular or multi-infarct dementia; mixed dementia; frontotemporal dementia such as Pick’s disease; and dementia related to stroke, Parkinson’s or Creutzfeldt-Jakob diseases)
Hemiplegia or Hemiparesis	Hemiplegia or Hemiparesis
Malnutrition	Malnutrition (protein or calorie) or at risk for malnutrition
Anxiety disorder	Anxiety disorder
Depression	Depression (other than bipolar)
Schizophrenia	Schizophrenia (e.g., schizoaffective and schizophreniform disorders)
PTSD	Post-traumatic stress disorder (PTSD)
Respiratory failure	Respiratory failure

Tab.2 List of 35 selected covariates with the associated descriptions

Tab.3 Summary of covariates identified by various feature selection methods

Tab.4 Prediction performance (C-index) comparison between the proposed and alternative models

Fig.1 Prediction performance (C-index) comparison under different discharge dispositions.

Covariate	$κ^j$	$S E (κ^j)$	p-value
Community discharge
ADL score	?0.113	0.014	1.11E-15	***
Mood score	?0.143	0.041	0.0005	***
Cancer	?0.436	0.162	0.0072	**
Anemia	?0.203	0.097	0.0367	*
Hypertension	?0.559	0.100	0	***
BPH	?0.515	0.141	0.0003	***
Renal disease	?0.330	0.134	0.0136	*
MDRO	?0.628	0.295	0.0331	*
Hip fracture	?0.604	0.233	0.0096	**
Other fracture	?0.421	0.127	0.0009	***
Non-Alzheimer’s dementia	?0.448	0.140	0.0014	**
Hemiplegia or Hemiparesis	?0.846	0.239	0.0004	***
Malnutrition	?0.556	0.196	0.0045	**
Re/Hospitalization
ADL score	0.087	0.024	0.00026	***
Anemia	0.482	0.180	0.00727	**
Obstructive uropathy	1.028	0.307	0.00080	***
Diabetes mellitus	0.503	0.170	0.00311	**

Tab.5 Significant covariates identified by the proposed model for each disposition

Fig.2 Marginal effect of ADL score on survival curves for community discharge.

Fig.3 Marginal effect of ADL score on survival curves for re/hospitalization.

Fig.4 Survival curves of a healthy resident under different discharge dispositions.

Fig.5 Survival curves of an unhealthy resident under different discharge dispositions.

Fig.6 Sampling algorithm prediction performance.

Fig.7 Discharge disposition prediction.

Fig.8 Comparison of prediction performance between the proposed sampling algorithm and alternative models for all discharge dispositions.

Fig.9 Comparison of prediction performance between the proposed sampling algorithm and alternative models for hospital transfer disposition.

Fig.10 Comparison of prediction performance between the proposed sampling algorithm and alternative models for community discharge disposition.

Tab.6 Experimental settings of acuity scenarios in the simulation study

Tab.7 Facility-level performance results under various census acuity scenarios of an NH

Fig.11 Comparison of facility-level performance results under various census acuity scenarios.

1	P C Austin, D M Rothwell, J V Tu, (2002). A comparison of statistical modeling strategies for analyzing length of stay after CABG surgery. Health Services and Outcomes Research Methodology, 3( 2): 107–133 https://doi.org/10.1023/A:1024260023851
2	P Cappanera, F Visintin, C Banditori, (2014). Comparing resource balancing criteria in master surgical scheduling: A combined optimisation-simulation approach. International Journal of Production Economics, 158: 179–196 https://doi.org/10.1016/j.ijpe.2014.08.002
3	K Carey, (2002). Hospital length of stay and cost: A multilevel modeling analysis. Health Services and Outcomes Research Methodology, 3( 1): 41–56 https://doi.org/10.1023/A:1021530924455
4	for Medicare CentersServices (CMS) Medicaid (2013). MDS 3.0 Quality Measures: User’s Manual. Research Triangle Park, NC: RTI International
5	Centers for Medicare and Medicaid Services (CMS) (2017). Long-Term Care Facility Resident Assessment Instrument 3.0 User’s Manual
6	S R Cole, H Chu, S Greenland, (2014). Maximum likelihood, profile likelihood, and penalized likelihood: A primer. American Journal of Epidemiology, 179( 2): 252–260 https://doi.org/10.1093/aje/kwt245
7	D R Cox, (1972). Regression models and life-tables. Journal of the Royal Statistical Society: Series B, Methodological, 34( 2): 187–202 https://doi.org/10.1111/j.2517-6161.1972.tb00899.x
8	R B D’Agostino, B H Nam, (2003). Evaluation of the performance of survival analysis models: Discrimination and calibration measures. Handbook of Statistics, 23: 1–25 https://doi.org/10.1016/S0169-7161(03)23001-7
9	S EikenK SredlL GoldJ KastenB BurwellP Saucier (2014). Medicaid Expenditures for Long-Term Services and Supports in FFY 2012. Bethesda, MD: Truven Health Analytics
10	E El-Darzi, C Vasilakis, T Chaussalet, P H Millard, (1998). A simulation modelling approach to evaluating length of stay, occupancy, emptiness and bed blocking in a hospital geriatric department. Health Care Management Science, 1( 2): 143–149 https://doi.org/10.1023/A:1019054921219
11	M Faddy, N Graves, A Pettitt, (2009). Modeling length of stay in hospital and other right skewed data: Comparison of phase-type, gamma and log-normal distributions. Value in Health, 12( 2): 309–314 https://doi.org/10.1111/j.1524-4733.2008.00421.x
12	S A Fashaw, K S Thomas, E McCreedy, V Mor, (2020). Thirty-year trends in nursing home composition and quality since the passage of the Omnibus Reconciliation Act. Journal of the American Medical Directors Association, 21( 2): 233–239 https://doi.org/10.1016/j.jamda.2019.07.004
13	R B GinsburgUS Supreme Court of the (1998). US Reports: Olmstead v. L.C., 527 US 581
14	D GonzàlezM PiñaL (2008) Torres. Estimation of parameters in Cox’s proportional hazard model: Comparisons between Evolutionary Algorithms and the Newton-Raphson Approach. In: Mexican International Conference on Artificial Intelligence. Berlin, Heidelberg: Springer, 513–523
15	P R Hachesu, M Ahmadi, S Alizadeh, F Sadoughi, (2013). Use of data mining techniques to determine and predict length of stay of cardiac patients. Healthcare Informatics Research, 19( 2): 121–129 https://doi.org/10.4258/hir.2013.19.2.121
16	F E Harrell Jr, K L Lee, D B Mark, (1996). Multivariable prognostic models: Issues in developing models, evaluating assumptions and adequacy, and measuring and reducing errors. Statistics in Medicine, 15( 4): 361–387 https://doi.org/10.1002/(SICI)1097-0258(19960229)15:4<361::AID-SIM168>3.0.CO;2-4
17	A A Holup, K Hyer, H Meng, L Volicer, (2017). Profile of nursing home residents admitted directly from home. Journal of the American Medical Directors Association, 18( 2): 131–137 https://doi.org/10.1016/j.jamda.2016.08.017
18	N R Hoot, L J LeBlanc, I Jones, S R Levin, C Zhou, C S Gadd, D Aronsky, (2008). Forecasting emergency department crowding: A discrete event simulation. Annals of Emergency Medicine, 52( 2): 116–125 https://doi.org/10.1016/j.annemergmed.2007.12.011
19	R A Incalzi, A Gemma, O Capparella, L Terranova, P Porcedda, E Tresalti, P Carbonin, (1992). Predicting mortality and length of stay of geriatric patients in an acute care general hospital. Journal of Gerontology, 47( 2): M35–M39 https://doi.org/10.1093/geronj/47.2.M35
20	A Kelly, J Conell-Price, K Covinsky, I S Cenzer, A Chang, W J Boscardin, A K Smith, (2010). Length of stay for older adults residing in nursing homes at the end of life. Journal of the American Geriatrics Society, 58( 9): 1701–1706 https://doi.org/10.1111/j.1532-5415.2010.03005.x
21	A A Kramer, J E Zimmerman, (2010). A predictive model for the early identification of patients at risk for a prolonged intensive care unit length of stay. BMC Medical Informatics and Decision Making, 10( 1): 27 https://doi.org/10.1186/1472-6947-10-27
22	L C McGuire, E S Ford, C A Okoro, (2007). Natural disasters and older US adults with disabilities: Implications for evacuation. Disasters, 31( 1): 49–56 https://doi.org/10.1111/j.1467-7717.2007.00339.x
23	M MoonB GageA Evans (1997). An examination of key Medicare provisions in the Balanced Budget Act of 1997. New York: The Commonwealth Fund
24	Y Murad, (2011). Skilled nursing facilities and post-acute care. Journal of Gerontology & Geriatric Research, 1( 101): 1–4
25	Investment Center for Seniors Housing & Care (NIC) National (2018). Skilled Nursing Data Report: Key Occupancy & Revenue Trends. 4Q2017
26	P W New, K Stockman, P A Cameron, J H Olver, J U Stoelwinder, (2015). Computer simulation of improvements in hospital length of stay for rehabilitation patients. Journal of Rehabilitation Medicine, 47( 5): 403–411 https://doi.org/10.2340/16501977-1957
27	P C Pendharkar, H Khurana, (2014). Machine learning techniques for predicting hospital length of stay in Pennsylvania federal and specialty hospitals. International Journal of Computer Science & Applications, 11( 3): 45–56
28	M Taboada, E Cabrera, M L Iglesias, F Epelde, E Luque, (2011). An agent-based decision support system for hospitals emergency departments. Procedia Computer Science, 4: 1870–1879 https://doi.org/10.1016/j.procs.2011.04.203
29	L Turgeman, J H May, R Sciulli, (2017). Insights from a machine learning model for predicting the hospital Length of Stay (LOS) at the time of admission. Expert Systems with Applications, 78: 376–385 https://doi.org/10.1016/j.eswa.2017.02.023
30	J Wang, J Li, K Tussey, K Ross, (2012). Reducing length of stay in emergency department: A simulation study at a community hospital. IEEE Transactions on Systems, Man, and Cybernetics: Part A, Systems and Humans, 42( 6): 1314–1322 https://doi.org/10.1109/TSMCA.2012.2210204
31	H Xie, T J Chaussalet, P H Millard, (2005). A continuous time Markov model for the length of stay of elderly people in institutional long-term care. Journal of the Royal Statistical Society, Series A (Statistics in Society), 168( 1): 51–61 https://doi.org/10.1111/j.1467-985X.2004.00335.x
32	X Zhang, S Barnes, B Golden, M Myers, P Smith, (2019). Lognormal-based mixture models for robust fitting of hospital length of stay distributions. Operations Research for Health Care, 22: 100184 https://doi.org/10.1016/j.orhc.2019.04.002

Viewed

Full text

Abstract

Cited

Shared

Discussed