PREDICTING PEOPLES’ FUTURE MOVEMENT BEHAVIOR PATTERN AT DISCHARGE FROM INPATIENT CARE AFTER STROKE: DEVELOPMENT OF A CLINICAL PREDICTION RULE

E.J.A. van der Neut-Piels^1,2,3, R. Wondergem⁴, M.W. Heymans⁵, A.A. van Kuijk⁶, J. Visser-Meily^7,2, C. Veenhof^8,3,2, M.F. Pisters^3,1,2

¹Fontys University of Applied Sciences, Exploring Healthy Behavior Research Group, Eindhoven, Netherlands, ²UMC Utrecht Brain Center, Utrecht University, Department of Rehabilitation, Physical Therapy Science and Sport, Utrecht, Netherlands, ³Center for Physical Therapy Research and Innovation in Primary Care, Julius Health Care Centre, Utrecht, Netherlands, ⁴Fontys University of Applied Sciences, School of Sport Studies, Eindhoven, Netherlands, ⁵VU University Medical Center, Department of Epidemiology and Data Science, Amsterdam, Netherlands, ⁶Tolbrug Rehabilitation Centre, Jeroen Bosch Hospital, 's Hertogenbosch, Netherlands, ⁷Centre of Excellence for Rehabilitation Medicine, UMC Utrecht Brain Centre, University Medical Centre Utrecht, and De Hoogstraat Rehabilitation, Utrecht, Netherlands, ⁸Research Group Innovation of Human Movement Care, University of Applied Sciences Utrecht, Utrecht, Netherlands

Background: Sedentary and inactive movement behavior (MB) patterns are a risk factor for recurrent cardiovascular events and a decline in physical functioning. Three MB patterns have been identified in people after stroke, one active (‘sedentary exercisers’) and two inactive patterns (sedentary movers’ or ‘sedentary prolongers’). An adequate selection of people with inactive MB patterns is required to stratify treatment. Insight into future predominant MB at home is needed in the clinical setting where treatment references occurs.

Purpose: The aim of this study was to develop and internaly validate a clinical prediction model to predict the long-term predominant MB pattern of people after stroke at discharge to the home setting.

Methods: Baseline factors measured within three weeks after discharge ofall 200 participants with first-ever stroke from the RISE prospective cohort study (2014-2019) and MB patterns at baseline, six months, one year, and two years were included to develop a clinical prediction rule. A two-step prediction model predicting the outcome measure, predominant MB pattern (either ‘sedentary exerciser’, ‘sedentary mover’ or ‘sedentary prolonger’) of 4 points in time over a two-year follow-up period . GEE analysis with backward selection was used to develop several variants of the two-step prediction model. The best performing and clinically easy-to-use model was selected and internally validated.

Results:​Female sex (B=-0,915 p=0,001), higher age (B=0,043) p=0,000) and feeling more fatigued (CIS-F) (B=-0,030 p=0,002) were found to be significant predictors to distinguish the sedentary exercisers from the inactive ‘sedentary movers’ and ‘sedentary prolongers’ MB patterns for home-based MB physiotherapy intervention. Discrimination (AUC) was 0,730. Walking speed (5MWT) (B=0,0737 p=0,005) in combination with MB pattern at discharge (B=2,461 p=0,000) could then discriminate excellently between the ‘sedentary prolongers’ and ‘sedentary movers’. (AUC = 0,827).

Conclusions: We developed a clinical prediction model to predict people's predominant 2-year MB pattern after a first ever stroke at discharge to the home setting. People’s MB can be accurately predicted with simple predictors following two clinical relevant steps. This enables clinicians to provide people required and stratified treatment.

Implications: People with a sedentary exercisers MB pattern can be easily excluded from further MB treatment at home withouth additional unnecesary measurement of their MB. Sedentary movers and sedentary prolongers can be refered to MB treatment at home and can then be distinguished adequately to stratify treatment.

Funding acknowledgements: This study was funded by the Netherlands Organization for Scientific Research (NWO), Doctoral grant for Teachers, 023.015.065, to ENP.

Keywords:
Stroke
Movement behavior
Treatment stratification

Topics:
Neurology: stroke
Health promotion & wellbeing/healthy ageing/physical activity
Non-communicable diseases (NCDs) & risk factors

Did this work require ethics approval? Yes

Institution: the Medical Ethics Research Committee of the University Medical Center Utrecht

Committee: the Medical Ethics Research Committee of the University Medical Center Utrecht

Ethics number: 14/76