Occupational Therapist Treatment of Patients in the Neurological Critical Care Unit: Utilization and Patient Characteristics

Method

Results

VIFs of independent variables in our analyses ranged from 1.0 to 1.5, which indicates weak relationships between independent variables and a low risk of multicollinearity.

Discussion

The current study illustrates the relationship between patient characteristics and occupational therapy service use in the NCCU. Our findings suggest that patients who have more comorbidities but who have a higher level of consciousness and who are in the NCCU for a longer time are more likely to receive occupational therapy services. The results confirm those of a study of critically ill survivors of traumatic brain injury (Chua et al., 2010), which found patients with a greater degree of illness were more likely to be seen by rehabilitation practitioners, were seen more frequently, and had a longer LOS. Previous clinical trials have shown no significant relationship between level of consciousness and rehabilitation potential (DeFina et al., 2010; Toschlog et al., 2003). These trials did not use decisions about which patients might have been the best candidates for rehabilitation on the basis of GCS score as part of their protocols—rather, all patients received services. In actual practice, as our study indicates, occupational therapists may need to make decisions about which patients to see in the NCCU on the basis of level of consciousness. Although it is not a sole determinant of rehabilitation use, NCCU occupational therapists likely consider level of consciousness, among other factors, when prioritizing patients for services.

Early onset of rehabilitation is generally associated with more favorable outcomes (Turon et al., 2018), as has been demonstrated by early mobilization (Fuest & Schaller, 2018) and neurocognitive protocols (Turon et al., 2018) implemented in critical care units. A higher level of consciousness was associated with earlier onset of occupational therapy—again, patients who are more alert will be able to participate in occupational therapy interventions sooner than those with a lower level of consciousness. Other research has supported this practical conclusion, showing that occupational therapy services were delayed for patients with reduced alertness secondary to sedation (Dinglas et al., 2013). A similar phenomenon may have occurred among some patients we studied, suggesting that NCCU occupational therapists and physicians collaborate to modify sedation protocols to enable earlier rehabilitation. Although in our model onset of occupational therapy was mainly a function of initial GCS, other important variables for which we were unable to account, such as severity of illness, may influence the timing of occupational therapy services and should be considered in future research.

Each additional day a patient spent in the NCCU was associated with a half-day later onset of occupational therapy services. Previous research has demonstrated that patients who are less healthy tend to be hospitalized longer (Howard et al., 2018). In a previous study, we found that physicians’ ratings of patients’ severity of illness were positively correlated with LOS and occupational therapy onset; that is, a greater level of illness was associated with greater NCCU LOS and longer occupational therapy onset (Malcolm et al., 2019). Extrapolating from the findings of our and others’ work, we believe that NCCU patients with greater severity of illness tend to be hospitalized longer and receive occupational therapy services later because they are unable to participate in therapy early in their NCCU stay.

Much attention has been paid to early mobility programs entailing in-bed exercises, standing, or walking (Corcoran et al., 2017), but little research about other occupational therapy interventions has been published. Underestimating a NCCU patient’s ability to participate in therapy might lead one to choose interventions in which the patient is passively engaged (i.e., passive range of motion, positioning). In fact, a study of physical therapy revealed the odds of receiving range-of-motion interventions in the NCCU was substantially higher than the odds of receiving other, more active interventions (Sottile et al., 2015). However, our findings reveal that passive interventions are the least billed category; instead, occupational therapists delivered significantly more interventions in the ADL–Home and Func–Cog intervention categories. Recent recognition that independence in functional activities, activities of daily living (ADLs), and instrumental activities of daily living (IADLs) are important outcomes suggests that occupational therapy interventions involving self-care and functional activities are necessary for critically ill patients (Weinreich et al., 2017).

Although our findings indicate that the likelihood of receiving ADL and IADL interventions increased with each additional occupational therapy session (as was true for all intervention categories), this likelihood decreased as LOS increased. Patients who stay in the NCCU longer—who are likely dealing with complex health issues (Howard et al., 2018)—may lack the necessary capacity to actively engage in ADL and IADL interventions. Alternatively, a clinician’s choice of intervention may be influenced by a perception that these patients are less able to actively engage in therapy sessions; prior research has suggested that severity of illness can affect occupational therapists’ clinical decision making in critical care (Dinglas et al., 2013). Related to this finding, receipt of passive therapeutic exercise interventions was more likely for patients who had a lower level of consciousness. Certainly, future research is needed to parse out the extent to which ADL and IADL interventions are possible in relation to NCCU patients’ LOS, health status, and level of consciousness.

Limitations

Our study had several limitations. Our findings come from a single NCCU at an academic hospital. Because wide variations exist in staffing and practices across critical care units, a future multisite study is needed to examine relationships between patient characteristics and occupational therapy use. Our analyses included number of comorbidities, but we were unable to account for the type and severity of comorbidities. Future research should include a comorbidity burden measure, such as the Elixhauser Comorbidity Index (Elixhauser et al., 1998), which provides a summary index score weighted by the severity of each comorbid condition (Moore et al., 2017). Some of our models—for example, the model predicting receipt of occupational therapy services—accounted for a modest amount of total variance.

Several important variables likely related to occupational therapy use were not available to us. For example, data on patients’ functional status, previous living situation, and other contextual and environmental factors may be important predictors of service provision in the NCCU. In addition, patients’ type of insurance was not available in our data. Although a 2006 study demonstrated differences in intensive care use related to insurance (Danis et al., 2006), more current research relating insurance and NCCU use is needed to account for shifting health care practices and policies, and other social determinants of health (e.g., socioeconomic status) should likewise be considered.

Both the receipt and onset of occupational therapy services may be influenced by referral processes, workflows, or other protocols physicians use in the NCCU. In our setting, variation exists in terms of the ordering of occupational therapy services. Certain diagnosis-specific order sets (e.g., with subarachnoid hemorrhage or any type of cerebrovascular accident) trigger an occupational therapy consult. Patients may otherwise receive occupational therapy services through recommendations that come from the multidisciplinary team during the provision of other services or from discussions occurring in multidisciplinary rounding of patients. Future research should consider administrative processes and physician preferences that influence whether and when occupational therapy services are initiated. We used billing data to ascertain broad categories of interventions delivered by occupational therapists. Because more granular intervention details are typically not mandatory fields in therapy flow sheets and are more commonly recorded in progress notes, future research may need to use natural-language processing techniques to obtain a more detailed picture of NCCU interventions. Finally, functional outcomes were not available for the patients in this study because documentation of therapy outcomes was primarily recorded in therapy progress notes.

Implications for Occupational Therapy Practice

This study has the following implications for occupational therapy practice and research:

Conclusion

This study helps to explain the relationships among patient factors, NCCU occupational therapy utilization, and occupational therapy interventions. Our findings suggest that the receipt and timing of occupational therapy in the NCCU depend on the patient’s comorbidities and level of consciousness. Prospective research is needed to tease out the influence of practitioner decision making regarding occupational therapy utilization and the types of interventions delivered to NCCU patients. As we consider the study findings within the context of the NCCU standards, we suggest that future research must refine these standards with respect to occupational therapy utilization and outcomes.