Medical Claims and Benefits Payable

The following table provides the details of our medical claims and benefits payable (including amounts payable for the provision of long-term services and supports, or LTSS) as of the dates indicated.



	December 31,
	2017		2016		2015
	(In millions)
Fee-for-service claims incurred but not paid (IBNP)	$	1,717	$	1,352	$	1,191
Pharmacy payable	112		112		88
Capitation payable	67		37		140
Other	296		428		266
	$	2,192	$	1,929	$	1,685

“Other” medical claims and benefits payable include amounts payable to certain providers for which we act as an intermediary on behalf of various government agencies without assuming financial risk. Such receipts and payments do not impact our consolidated statements of operations. Non-risk provider payables amounted to $122 million, $225 million and $167 million, as of December 31, 2017, 2016 and 2015, respectively.

The following table presents the components of the change in our medical claims and benefits payable for the periods indicated. The amounts presented for “Components of medical care costs related to: Prior periods” represent the amount by which our original estimate of medical claims and benefits payable at the beginning of the period were less (more) than the actual amount of the liability based on information (principally the payment of claims) developed since that liability was first reported.



	Year Ended December 31,
	2017			2016			2015
	(In millions)
Medical claims and benefits payable, beginning balance	$	1,929		$	1,685		$	1,201
Components of medical care costs related to:
Current period	17,037			14,966			11,935
Prior periods	36			(192		)	(141		)
Total medical care costs	17,073			14,774			11,794

Change in non-risk provider payables	(106		)	58			48

Payments for medical care costs related to:
Current period	15,130			13,304			10,448
Prior periods	1,574			1,284			910
Total paid	16,704			14,588			11,358
Medical claims and benefits payable, ending balance	$	2,192		$	1,929		$	1,685

Reinsurance recoverables of $16 million, $61 million, and $46 million, as of December 31, 2017, 2016, and 2015, respectively, are included in “Receivables” in the accompanying consolidated balance sheets.

The following tables provide information about incurred and paid claims development as of December 31, 2017, as well as cumulative claims frequency and the total of incurred but not paid claims liabilities. The cumulative claim frequency is measured by claim event, and includes claims covered under capitated arrangements.



Incurred Claims and Allocated Claims Adjustment Expenses							Total IBNP		Cumulative number of reported claims
Benefit Year	2015 ⁽¹⁾		2016		2017		Total IBNP		Cumulative number of reported claims

	(In millions)
2015	$	12,113	$	11,928	$	11,939	$	6	84
2016			15,064		15,093		46		109
2017					17,037		1,665		114
					$	44,069	$	1,717



Cumulative Paid Claims and Allocated Claims Adjustment Expenses
Benefit Year	2015 ⁽¹⁾		2016		2017

	(In millions)
2015	$	10,615	$	11,906	$	11,932
2016			13,403		14,952
2017					15,130
					$	42,014

The following table represents a reconciliation of claims development to the aggregate carrying amount of the liability for medical claims and benefits payable.



	2017
	(In millions)
Incurred claims and allocated claims adjustment expenses	$	44,069
Less: cumulative paid clams and allocated claims adjustment expenses	(42,014		)
Non-risk provider payables and other	137
Medical claims and benefits payable	$	2,192

_______________________________


(1)	Data presented for this calendar year is required supplementary information, which is unaudited.

That portion of our total medical claims and benefits payable liability that is most subject to variability in the estimate is IBNP. Our IBNP, as included in medical claims and benefits payable, represents our best estimate of the total amount of claims we will ultimately pay with respect to claims that we have incurred as of the balance sheet date. We estimate our IBNP monthly using actuarial methods based on a number of factors.

Assuming that our initial estimate of IBNP is accurate, we believe that amounts ultimately paid would generally be between 8% and 10% less than the IBNP liability recorded at the end of the period as a result of the inclusion in that liability of the provision for adverse claims development and the accrued cost of settling those claims. Because the amount of our initial liability is an estimate (and therefore not perfectly accurate), we will always experience variability in that estimate as new information becomes available with the passage of time. Therefore, there can be no assurance that amounts ultimately paid out will fall within the range of 8% to 10% lower than the liability that was initially recorded. Indeed, the amount we paid out during 2017 in satisfaction of our liability for medical claims and benefits payable at December 31, 2016, was in excess of the amount originally accrued. Furthermore, because our initial estimate of IBNP is derived from many factors, some of which are qualitative in nature rather than quantitative, we are seldom able to assign specific values to the reasons for a change in estimate—we only know when the circumstances for any one or more factors are out of the ordinary.

The use of a consistent methodology (including a consistent provision for adverse claims development) in estimating our liability for medical claims and benefits payable minimizes the degree to which the overestimation of that liability at the close of one period may affect consolidated results of operations in subsequent periods. In particular, the use of a consistent methodology should result in the replenishment of reserves during any given period in a manner that generally offsets the benefit of favorable prior period development in that period.

Facts and circumstances unique to the estimation process at any single date, however, may still lead to a material impact on consolidated results of operations in subsequent periods. For example, any absence of adverse claims development (as well as the expensing through general and administrative expense of the costs to settle claims held at the start of the period) will lead to the recognition of a benefit from prior period claims development in the period subsequent to the date of the original estimate, to the extent that replenishment of reserves is not equal to the benefit recognized due to the absence of adverse development.

Conversely, in the presence of adverse claims development, the financial impact of recording claims expense in the current period that is related to dates of service in the prior period will be compounded by the need to replenish the provision for adverse development.

As noted above, the amounts ultimately paid out in 2017 for dates of service in 2016 and prior were in excess of the liability we had established for medical and claims and benefits payable at December 31, 2016. In contrast, the amounts we paid out for prior period dates of service in 2016 and 2015 were less than the liabilities we had established at the end of the previous years. In all three years, the differences between our original estimates and the amounts ultimately paid out for the most part related to IBNP. While many related factors working in conjunction with one another serve to determine the accuracy of our estimates, we are seldom able to quantify the impact that any single factor has on a change in estimate. In addition, given the variability inherent in the reserving process, we will only be able to identify specific factors if they represent a significant departure from expectations. As a result, we do not expect to be able to fully quantify the impact of individual factors on changes in estimates.

2017

We recognized unfavorable prior period claims development in the amount of $36 million for the year ended December 31, 2017. This amount represented the extent to which our initial estimate of medical claims and benefits payable at December 31, 2016, was less than the amount that was ultimately paid out in satisfaction of that liability; but does not include additional amounts expensed to re-establish the margin for adverse development. We believe this underestimation was due primarily to the following:


•	Inaccurate adjudication of provider claims at our Florida, Illinois, New Mexico and Puerto Rico health plans that created substantial payment backlogs which we were unable to adequately measure when we estimated our liability at December 31, 2016.

We believe that the most significant uncertainties relating to our estimated IBNP liability at December 31, 2017, are as follows:


•	At our Florida health plan, the inventory of unpaid claims increased significantly during the first two quarters of 2017, then dropped in the third quarter and fourth quarters of 2017. Changes in claim inventories impact the timing between dates of service and the dates claims are paid, making our liability estimates subject to more than the usual amount of uncertainty.


•	At our Illinois health plan, in 2017 we paid a large number of claims that had previously been denied and were subsequently disputed by providers. We have also established a liability for additional expected claims resulting from provider disputes. This has created some distortion in the claims payment patterns, making our liability estimates subject to more than the usual amount of uncertainty.


•	At our California health plan, adjustments to our inpatient authorization process have improved the timeliness of paying inpatient claims, making our liability estimates subject to more than the usual amount of uncertainty.

•

In 2017 we implemented a new process for increased quality review of claims payments in six of our health plans. While we do not anticipate this new process will impact the percentage of claims paid within the timely turnaround requirements, we believe it will have a minor impact on the timing of some paid claims. For this reason, our liability estimates in the six health plans are subject to more than the usual amount of uncertainty.


•	At our Puerto Rico health plan, Hurricane Maria had a significant impact on both utilization of services and our ability to process claims payments in Puerto Rico. For these reasons, we believe our liability estimates are subject to more than the usual amount of uncertainty.


•	December 2017 data from the Centers for Disease Control and Prevention have indicated widespread influenza activity in several states in which we operate health plans. Although we have established liabilities for additional expected claims related to influenza, our liability estimates are subject to more than the usual amount of uncertainty.

2016

We recognized favorable prior period claims development in the amount of $192 million for the year ended December 31, 2016. This amount represented the extent to which our initial estimate of medical claims and benefits payable at December 31, 2015, was more than the amount that was ultimately paid out in satisfaction of that liability. We believe this overestimation was due primarily to the following:

•

A new version of diagnostic codes was required for all claims with dates of service on October 1, 2015, and later. As a result, payment was delayed or denied for a significant number of claims due to provider submission of claims with diagnostic codes that were no longer valid. Once providers were able to submit claims with the correct diagnostic codes, our actual costs were ultimately less than expected.


•	At our New Mexico health plan, we overestimated the impact of several pending high-dollar claims, and our actual costs were ultimately less than expected.


•	At our Washington health plan, we overpaid certain outpatient facility claims in 2015 when the state converted to a new payment methodology. We did not include an estimate in our December 31, 2015, reserves for this potential recovery.


•	At our California health plan, we enrolled approximately 55,000 new Medicaid Expansion members in 2015. For these new members, our actual costs were ultimately less than expected.

2015

We recognized favorable prior period claims development in the amount of $141 million for the year ended December 31, 2015. This amount represented the extent to which our initial estimate of medical claims and benefits payable at December 31, 2014 was more than the amount that was ultimately paid out in satisfaction of that liability. We believe this overestimation was due primarily to the following:

•

At our Ohio and California health plans, approximately 61,000 and 100,000 members, respectively, were enrolled in the new Medicaid expansion program during 2014. Also in Ohio, approximately 17,000 members were enrolled in the new MMP program in 2014. Because we lacked sufficient historical claims data, we initially estimated the reserves for these new members based upon a number of factors that included pricing assumptions provided by the state; our expectations regarding pent up demand; our beliefs about the speed at which new members would utilize health care services; and other factors. Our actual costs were ultimately less than expected.

•

At our New Mexico health plan, the state implemented a retroactive increase to the provider fee schedules in mid-2014. As a result, many claims that were previously settled were reopened, and subject to, additional payment. Because our reserving methodology is most accurate when claims payment patterns are consistent and predictable, the payment of additional amounts on claims that in some cases had been settled more than six months before added a substantial degree of complexity to our liability estimation process. Due to the difficulties in addressing that added complexity, liabilities recorded as of December 31, 2014 were in excess of amounts ultimately paid.


•	At our Washington health plan, we collected amounts in 2015 related to certain claims paid in 2013. Such collections were not anticipated in our reserves as of December 31, 2014.