Date Presented 3/31/2017
We examined payments for Medicare patients with stroke before and after implementation of the new coding system to illustrate how to use propensity score matching approaches to measure the impact of payment interventions while avoiding bias due to differences in patient mix or patient severity.
Primary Author and Speaker: Ickpyo Hong
Additional Authors and Speakers: Annie Simpson, Craig Velozo, Moon Young Kim, Kit Simpson
PURPOSE: The purpose of this study was to examine rehabilitation service payments for patients with stroke in 2012, before the new billing code rules were implemented, and in 2013, during the fiscal year following the policy change. The main hypothesis was that payments for rehabilitation services would change if clinicians were required to document patients’ functional status using the new functional measurement coding system.
BACKGROUND: As part of reforming the Medicare payment system for outpatient therapy services, the Centers for Medicare & Medicaid Services (CMS) mandated the use of nonpayable G-codes and seven modifiers for reimbursement since January 1, 2013 (CMS, 2012). CMS reasoned that if clinicians assessed patients’ functional status in a timely manner, then patients would receive only necessary interventions, and use of extended and unnecessary interventions would be reduced; thus, the overall rehabilitation payment for patients would decrease. We assessed whether or not the new coding system affected patients’ overall rehabilitation payment, controlling for effects of differences in stroke severity case mix and comorbid conditions.
DESIGN: This was a quasi-experimental retrospective secondary data analysis study comparing matched groups of patients from a 5% sample of 2012 and 2013 Medicare claim data. Participants were 4,765 patients with stroke identified from the two data sets (n = 2,356 in 2012 and n = 2,409 in 2013). Patients had a mean age of 77.9 yr (SD = 7.1). The majority of patients were female (n = 2,588, 54.3%) and White (82.6%) and had ischemic stroke (90.1%). The stroke severity of the patients included mild (35.1%), moderate (18.3%), and severe (46.6%).
METHOD: Patients with stroke who received rehabilitation services (Medicare Part B benefits) were extracted from the 2012 and 2013 Medicare claim data sets. Between 2012 and 2013, rehabilitation costs per person and average costs per visit were estimated using a multivariable log-linked gamma-distributed generalized linear model with inverse probability of treatment weighting (Rosenbaum, 2002). We investigated the differences between the two groups’ demographics and comorbidities.
RESULTS: Except for two comorbidities (disorders of lipoid metabolism and speech disturbance), we found no differences in demographic characteristics and comorbidities (all, p > .05) between the 2012 and the 2013 stroke cohorts. Medicare’s rehabilitation costs per person and average costs per visit in 2013 were significantly less than the payments in 2012 (p < .05). Average rehabilitation costs per person were $4,220 in 2012 and $3,813 in 2013. Average costs per visit were $823 in 2012 and $745 in 2013.
CONCLUSION: After we implemented G-code systems in rehabilitation outpatient settings, the 1-yr cost to Medicare for outpatient stroke treatment decreased from 2012 to 2013. The new coding system could allow clinicians to measure patients’ functional status in a timely manner and avoid unneeded extended interventions. However, it is not clear if the lack of practitioner coding experience resulted in undercoding and a high rate of claims refusals, leading to change in use of treatments that were needed but difficult to get reimbursed. Further studies are needed to investigate the effects of the new coding systems on the outcomes of rehabilitation services, such as performance capacity for activities of daily living.
IMPACT STATEMENT: The new functional limitation coding system could allow clinicians to measure patients’ functional status in a timely manner and avoid extended interventions. To verify the study results, future studies need to compare patients’ functional outcome levels before and after the policy change.
References
Centers for Medicare & Medicaid Services. (2012). Medicare program; Revisions to payment policies under the physician fee schedule, DME face-to-face encounters, elimination of the requirement for termination of non-random prepayment complex medical review and other revisions to Part B for CY 2013
; Final rule. Retrieved from https://www.gpo.gov/fdsys/pkg/FR-2012-11-16/pdf/2012-26900.pdf
Rosenbaum, P. R. (2002). Observational studies. New York: Springer.
