Collaborative Targeted Case Management in Improving Functional Status in Patients With Stage III-IV Cancer

The mean change in the Ambulatory Post-Acute Care Basic Mobility Computer Adaptive Test (AM PAC CAT) score (month 6 minus baseline) and the [95% confidence interval] for patients assigned to each arm regardless of compliance are reported below. The AM PAC CAT is an item response theory-modeled scale that ranges from 0 to infinity. However, the score range of interest for the trial that corresponds to the ability to ambulate and transfer safely and independently (+/- gait aid) is 55 - 66 with higher scores corresponding to better mobility and the MID for improvement being 1.0. For the primary outcome group differences will be assessed over the 6 months of the trial using mixed-effects model repeated measures (MMRM) analysis which will include a random effect for patient, a fixed effect for measurement point (3 months or 6 months), and adjustment for age, gender, baseline value of the outcome variable, cancer type, cancer stage (IV versus other), and time.

The change in the Brief Pain Inventory (BPI) average score and total interference score (Month 6 minus baseline) and the and the [95% confidence interval] for patients assigned to each arm regardless of compliance. The BPI average is a single numerical rating scale (NRS) anchored at 0 and 11. The total inference score is a composite of 6 NRS scales. Higher scores indicate more pain or a worse health state. For both measures the minimally important difference (MID) is population dependent and ranges between 0.5 and 2.0. For the secondary outcome group differences will be assessed over the 6 months of the trial using mixed-effects model repeated measures (MMRM) analysis which will include a random effect for patient, a fixed effect for measurement point (3 months or 6 months), and adjustment for age, gender, baseline value of the outcome variable, cancer type, cancer stage (IV versus other), and time.

The change in health utility, as measured by the change in EuroQol 5-D (5Q-5D) score from baseline score (Month 6 minus baseline) and the and the [95% confidence interval] for patients assigned to each arm regardless of compliance. The EuroQol 5-D utility ranges from 0 to 1.0 with high values indicating better health states. The minimally important difference (MID) is population dependent and range between 0.5 and 1.0. For the secondary outcome group differences will be assessed over the 6 months of the trial using mixed-effects model repeated measures (MMRM) analysis which will include a random effect for patient, a fixed effect for measurement point (3 months or 6 months), and adjustment for age, gender, baseline value of the outcome variable, cancer type, cancer stage (IV versus other), and time.

The count of hospital admission lasting >24 hours. To assess the impact on number of hospitalizations we will use a negative binomial model, with an offset for number of days of follow-up.

The count of days spent in the hospital for admissions lasting >24 hours. For patients who had at least one hospitalization, the impact of the interventions on the total number of days in hospital will be assessed using mixed effect Poisson analysis.

Binary variable describing whether patients were discharged to an inpatient facility (skilled nursing facility, inpatient rehabilitation facility, hospice or long term acute care facility) or home with/without services. Logistic models with random effects for patient will be used to analyze inter-group differences.

Binary variable describing whether an admission was planned for anti-cancer treatment or unplanned. Logistic models with random effects for patient will be used to analyze inter-group differences.