Medicare’s Reimbursement Reduction for Nerve Conduction Studies: Effect on Use and Payments

To decrease health care costs, Medicare sought to identify overvalued Current Procedural Terminology (CPT) codes, including those with rapid volume growth, those submitted multiple times, or those submitted in conjunction with other codes.¹ The codes for nerve conduction studies (NCS) met all 3 criteria.² Electromyography (EMG) should typically be performed with NCS.³ Furthermore, EMG and NCS are part of the core residency training for neurologists and physiatrists but not for other health care professionals (physicians, podiatrists, physical therapists, nurse practitioners, and physician assistants). Medicare implemented a sharp reduction in reimbursement for NCS on January 1, 2013,⁴ but reimbursement for EMG was not changed. The effect of this policy on providers’ use of services and reimbursement payments for this is unclear.

Methods

A retrospective analysis was done between July and December 2015 of Medicare EMG and NCS use and payments to health care professionals for 2012 and 2013 using the Physician and Other Supplier Public Use File. Individual billable services were identified by CPT codes and G-codes. For neurologists, other commonly performed services were also investigated (Table 1).

Table 1.

Changes in Medicare Services Provided by 11 800 Neurologists and Resulting Payments From 2012 to 2013

Characteristic	2012			2013			Absolute Change in Payment per Neurologist 2012–2013, $
	Total Studies, No.	Proportion of Neurologists Performing Service, %	Mean Payment per Neurologist, $a	Total Studies, No.	Proportion of Neurologists Performing Service, %	Mean Payment per Neurologist, $a
All	12 295 597	100	96 803	11 611 844	100	87 706	−9097
Nerve conduction studies	465 796	46.6	15 260	418 728	39.8	5118	−10 142
Electromyography complete	371 834	40.1	3086	383 193	40.5	3146	60
Electromyography limited	76 288	14.2	311	77 493	14.0	324	13
Electromyography or nerve conduction study other	16 400	2.4	71	23 614	2.8	123	52
Evaluation and management	8 379 260	97.6	57 497	8 394 076	97.6	57 668	171
Electroencephalography	719 713	49.5	8166	717 041	49.0	9499	1333
Imaging	304 628	13.9	3193	299 422	13.5	3179	−14
Polysomnogram	132 468	9.1	2542	131 215	9.0	2411	−131
Injection or infusion	321 029	15.0	2549	324 250	15.4	2458	−91
Intraoperative monitoring	54 914	4.4	1015	44 604	4.0	508	−507
Evoked potentials	77 225	5.4	428	80 632	5.7	460	32
Neurostimulation	32 591	3.4	349	36 329	3.8	400	51
Vestibular	64 627	3.0	337	79 060	3.4	366	29
Neuropsychiatric	52 327	5.0	334	57 551	5.2	372	38
Autonomic	17 449	1.3	108	19 125	1.5	206	98
Other	535 621	18.1	1557	525 511	18.6	1469	−88

^aMean payment per neurologist represents the average payment among all neurologists in the sample, and not only among those who performed the relevant service.

To determine the number of visits for NCS in 2012, the highest number of visits for CPT codes 95900, 95903, and 95904 was identified. To determine the number of visits for NCS in 2013, the number of unique services for codes 95907 to 95913 was summed. A complete EMG was identified with codes 95860 to 95864 and 95886, and limited EMG with codes 95870 and 95885. Payments were calculated by multiplying the number of services by the average Medicare payment received. Analyses were limited to providers receiving a Medicare payment for any CPT code or G-code in both 2012 and 2013. Sensitivity analyses done with multilevel linear regression to adjust for average Hierarchical Condition Category score, hospital referral region, and percentage of EMG and/or NCS services performed at facilities did not change the inferences used in the analysis. Because this study involved research using publicly available data, it was not considered to be regulated by the University of Michigan Institutional Review Board.

Results

In 2012, a total of 11 336 health care professionals performed EMG and/or NCS, as compared with 9807 in 2013 (Table 2). Decreases in the numbers of NCS performed by neurologists were 47 068 (10.1%); by physiatrists, 25 366 (13.8%); and by other health care professionals, 46 676 (30.7%). The number of EMGs performed by neurologists and physiatrists changed by less than 3% from 2012 to 2013. In contrast, the number of EMGs performed by other health care professionals decreased by 3849 (7.3%). Of the other health care professionals performing EMG and/or NCS in 2012, a total of 41% stopped performing EMG and/or NCS in 2013. The number of nerves per NCS and limbs per EMG study changed by less than 7% for all health care professionals.

Table 2.

Changes in Medicare Use and Payments for Nerve Conduction Studies and Electromyography by All Medicare Provider Types From 2012 to 2013^a

Test	2012			2013			Relative Change in Total Studies 2012–2013, %
	Total Studies, No.	Study Length, Unit, Mean (SD)b	Mean Payment per Study, US $	Total Studies, No.	Study Length, Unit, Mean (SD)b	Mean Payment per Study, US $
Nerve Conduction Studies
Neurology	465 796	8.11 (2.81)	387	418 728	7.89 (2.63)	144	−10.1
Physiatry	183 665	7.32 (2.41)	360	158 299	7.45 (2.42)	140	−13.8
Other	151 756	9.23 (3.93)	485	105 080	9.87 (3.63)	176	−30.7
Total	801 217	8.21 (3.13)	398	682 107	8.19 (2.94)	148	−14.9
Electromyography Complete
Neurology	371 834	1.54 (0.45)	98	383 193	1.54 (0.44)	97	3.1
Physiatry	151 893	1.50 (0.38)	96	157 404	1.47 (0.34)	94	3.6
Other	46 270	1.76 (0.65)	105	43 028	1.70 (0.54)	108	−7.0
Total	569 997	1.56 (0.47)	98	583 625	1.54 (0.43)	97	2.4
Electromyography Limited
Neurology	76 288	1.42 (0.41)	48	77 493	1.44 (0.41)	49	1.6
Physiatry	18 646	1.37 (0.38)	58	17 864	1.38 (0.39)	60	−4.2
Other	6373	1.45 (0.48)	57	5766	1.50 (0.54)	57	−9.5
Total	101 307	1.41 (0.41)	51	101 123	1.43 (0.42)	52	−0.2

^aMedicare provider types include physicians, podiatrists, physical therapists, nurse practitioners, and physician assistants.

^bMean study length is the mean number of nerves (nerve conduction study) or limbs (complete or limited electromyography) as determined on the basis of corresponding current procedural terminology codes.

Total NCS payments to all health care professionals decreased by 68% ($219 million), from $319 653 039 to $101 108 830, with 22% ($48 million) attributable to a decrease in use. Health care professionals other than neurologists and physiatrists accounted for 47% of the decrease in payments through reductions in the use of NCS, even though they accounted for only 23% ($73 559 201 of $319 653 039) of the total payments for NCS in 2012. Total EMG payments to all health care professionals increased by $0.7 million (1.2%).

For the 11 800 neurologists included in the analysis, little change in the number of other commonly performed services (Table 1) was observed.

Discussion

The Medicare NCS reimbursement policy implemented on January 1, 2013, was associated with a 15% decrease in NCS use, from 801 217 to 682 107 studies, and with the $219 million decrease cited above in Medicare payments during the subsequent year. Other health care professionals disproportionately decreased their use of NCS relative to its use by neurologists and physiatrists, accounting for most of the savings from its decreased use. The use of EMG by neurologists and physiatrists changed little, whereas a decrease in its use among other health care providers was observed, attributable to a large proportion of these providers who stopped EMG and NCS altogether. The reduction in use of EMG and NCS by other health care professionals may be a positive outcome of the change in the Medicare reimbursement policy because these providers typically lack certification and residency training for EMG and NCS. Because NCS should usually be performed with EMG,³ the reduction in neurologists’ and physiatrists’ use of NCS without a concomitant decline in the use of EMG may reflect a decrease in unnecessary tests. Health care professionals also did not change the study length of NCS despite changing incentives for the number of nerves studied. Furthermore, neurologists did not increase their use of other services to compensate for the loss in NCS-related revenue. Past studies revealed that decreases in Medicare reimbursement did not reduce appropriate testing and treatment, but curtailed inappropriate treatment.^5,6 Although our study did not contain information on study appropriateness, the pattern of change in use of EMG and NCS suggests findings similar to those in past studies of Medicare reimbursement with regard to reducing inappropriate,butnotappropriate,testingandtreatment.Notably,ourdata cannot address the fairness of current NCS reimbursement. Other limitations of our study include the absence of technical payment information for EMG and NCS provided in a facility.