The cohort study was conducted in Taiwan, with investigators also observing changes over time in the influence of sex, stroke type, and copayment exemption type of rehabilitation utilization.
Utilization of poststroke rehabilitation was affected by geographic factors such as urban-rural gaps, as well as personal factors, a recent retrospective cohort study in Taiwan found.1 Additionally, the number of therapists was significantly associated with utilization of poststroke care, and influence of sex, stroke type, and copayment exemption type changed over time.
Investigators, including corresponding author Nicole Huang, PhD, adjunct professor, Institute of Hospital and Health Care Administration, National Yang Ming Chiao Tung University, in Taipei, Taiwan, included a total of 14,600 patients with stroke between 2005 and 2011. The primary outcome to be observed was utilization of physical therapy (PT) or occupational therapy (OT) at different periods after stroke onset.
Huang et al found that, of all patients, 44.7% received PT and 33.8% received OT in the 2-year period. Odds of receiving PT and OT were increased for patients with more severe stroke and more comorbid diseases, while odds decreased for patients who were older in age.
Patients with higher stroke severity index (SSI) scores had higher odds of receiving PT between 0 and 6 months (1.07), 7 and 12 months (1.08), 13 and 18 months (1.07), and 19 and 24 months (1.06; P <.05 for all). Patients with higher Charlson Comorbidity Index (CCI) scores also had higher odds of receiving PT between 0 and 6 months (1.30), 7 and 12 months (1.26), 13 and 18 months (1.24), and 19 and 24 months (1.21; P <.05 for all). These patients were also at higher odds of receiving OT between 0 and 6 months (SSI, 1.06; CCI, 1.28), 7 and 12 months (SSI, 1.08; CCI, 1.25), 13 and 18 months (1.07; CCI, 1.24), and 19 and 24 months (SSI, 1.07; CCI, 1.22; P <.05 for all).
Sex, stroke type, and copayment status influenced use of rehabilitation services during specific periods, and men had higher odds of receiving both PT and OT than women during the first year post stroke. Those who had nonischemic stroke had higher odds of receiving PT in the first 18 months, as well as OT in the first 6 months, with these odds “disappearing” in later periods, investigators noted. They also acknowledged the interesting finding related to copayment exemption, which was associated with lower odds of rehabilitation utilization in the first 6 months after admission. Compared with those who are not exempt from copayments, in later periods, exemption was associated with greater odds of receiving these rehabilitation services.
Availability of rehabilitation resources and urbanization were both significantly associated with odds of rehabilitation use, with those in rural and suburban areas having lower odds of use when compared with urban patients during the 2-year period. For rural patients, ORs were 0.82, 0.80, 0.80, and 0.71 between 0 and 6 months, 7 and 12 months, 13 and 18 months, and 19 and 24 months, respectively (P <.05 for all), compared with urban patients, where ORs were 0.91, 0.89, 0.85, and 0.86 for those same periods, respectively (P <.05 for all).
Investigators observed a similar trend when looking at OT, but patients in suburban areas differed in that they had higher odds of using OT in the first 6 months following stroke (OR, 1.16; P = .001) when compared with those in urban areas. Higher odds for both PT and OT use were noted for those living in areas with higher availability of these resources, with the strongest association observed in the first 6 months after stroke onset (PT: OR, 1.076, P = .024; OT: OR, 1.414, P <.001).
The study was limited by the claims data, which lack information on severity degree, which Huang et al addressed by using SSI to measure stroke severity. Access to outpatient rehabilitation was also noted, as investigators were only able to speculate on transportation barriers using urbanization of subjects’ residences. The availability of caregivers as a potential factor was not considered, and functional outcomes were unable to be evaluated due to the study being based on NHI secondary data. Lastly, the investigators noted that the location of a hospital may not be a “perfect indicator” for the precise location of a participant’s residence.