Recent Trends in Workers Compensation Coverage by Brian Z. Brown, FCAS Melodee J. Saunders, ACAS

Transcription

1 Recent Trends in Workers Compensation Coverage by Brian Z. Brown, FCAS Meodee J. Saunders, ACAS

2 TITLE: RECENT TRENDS IN WORKERS COMPENSATION COVERAGE BY: Ms. Meodee J. Saunders, A.C.A.S., M.A.A.A. Mr. Brian Z. Brown, F.C.A.S., M.A.A.A. Meodee J. Saunders is Vice President of Actuaria Services for Midwest Empoyers Casuaty Company in St. Louis, Missouri. Ms. Saunders hods a bacheor of science degree in mathematics from the University of Nebraska. She is an Associate of the Casuaty Actuaria Society and a Member of the American Academy of Actuaries. Ms. Saunders area of expertise is workers compensation insurance, particuary with respect to excess insurance for sef-insured empoyers. Brian Z. Brown is a Feow of the Casuaty Actuaria Society, a Member of the American Academy of Actuaries and a Consuting Actuary in the Miwaukee office of Miiman & Robertson, Inc. Brian is aso chairman of the CLRS Committee and hods a bacheor of science degree in Economics from Iinois State University. Brian has severa workers compensation cients who are using managed care concepts.

3 RECENT TRENDS IN WORJLERS COMPENSATION COVERAGE ABSTRACT As a ine of business, workers compensation has undergone many significant changes in the ast few years. Key eements at the forefront of change incude the foowing: Increased eves of retained exposure by empoyers; Rapid growth in managed care initiatives; and State enactment of comprehensive system reforms. Due to the above changes, actuaries invoved in reserving workers compensation coverage wi fmd it necessary to use new methodoogies and assumptions to correcty estimate reserve eves because historica oss data may not accuratey predict future cost eves and trends. When empoyers purchase arge deductibe insurance they retain the smaer more stabe osses and eave the catastrophic exposures to the insurer. This creates increased severities, decreased frequencies and onger taied reporting and payment patterns. Use of managed care techniques shoud decrease medica severities and shoud aso decrease indemnity severities and wi ikey cause a shift in frequency among types of injuries. The impact of statutory benefit eve reforms must be assessed before a affected caims are reported and setted. Thus the chaenge wi be to make we informed judgments as to the impact of such comprehensive changes on future reserve eves. The purpose of this paper is not so much to answer questions but rather to raise the types of questions the reserving actuary must ask in order to revise and revamp his or her approach to reserving workers compensation exposures, 3

4 RECENT TRENDS IN WORKERS COMPENSATION COVERAGE ODUCm Severa changes have occurred in the workers compensation marketpace in recent years. Three of the most significant changes which affect reserve eves are: Increased eves of retained exposure by empoyers; Rapid growth in managed care initiatives; and State enactment of comprehensive system reforms. The first section of this paper describes how the increased retention of the exposure by insureds affects standard reserving techniques. This section aso describes a reativey simpe modification to the standard Bornhuetter-Ferguson procedure which is used in reserving excessreinsurance products. The second section discusses managed care initiatives and how they impact standard reserving assumptions. This section aso provides a genera discussion of the various roes of the insurance carrier, empoyer, empoyee, case manager, caim adjuster and medica provider in a managed care setting. In describing these roes we outine the managed care process and highight some of the savings associated with managed care initiatives. The third section discusses some specia financia arrangements between insurance carriers and managed care organizations and discusses the effect these arrangements may have on reserve 4

5 eves. As a natura extension of this section, we outine some techniques for measuring managed care savings in workers compensation. The fourth section discusses how heath insurance principes can be used to derive a capitated rate for workers compensation medica costs. Finay, we take a brief ook at the types of workers compensation reforms that have occurred over the ast few years and how these reforms may affect oss reserves. I. HIGH Starting in the eary 1990 s, many carriers began to offer high deductibe poicies to their workers compensation insureds. These deductibes woud usuay range between $50,000 and $,OOO,OOO per occurrence. These products were offered to: Reduce the carriers share of the highy unprofitabe residua markets in severa states; Compete with sef-insurance and excess workers compensation products; Have the insured share in its own oss experience and directy benefit from effective risk management procedures and pay for ineffective procedures; and Market a product that fits in with some companies strategic pans.

6 These poicies create compications for many reserving anaysts who previousy may have ony reserved first doar workers compensation products. The extended and sow reporting patterns dispayed by many workers compensation industry statistics is amost unfathomabe. For exampe, recent data pubished by the Reinsurance Association of America impies that ony 50% of the osses are reported 8 years after the beginning of the accident year. If a primary company begins to write excess/high deductibe workers compensation products and does not separatey anayze this experience, reserve projection methods may produce biased resuts. We wi iustrate this through an exampe where the reserve anayst uses a simpe incurred oss projection method. However, instead of anayzing the data separatey for high deductibe products and primary products, the anayst assumes that the combined oss experience wi be refected in deveopment factors and resut in unbiased projections. This approach wi significanty understate a company s estimated reserves. To iustrate this point, assume: ) 1995 edition of Reinsurance Association of America. The 8 year period assumes a reativey ow per occurrence retention (e.g., $50, ,000). It woud take onger than 8 years for one-haf of the osses to be reported if the retention were higher. We woud aso note that reporting patterns differ significanty from company to company and some carriers (especiay those who speciaize in excess/high deductibe workers compensation exposures) may dispay significanty quicker reporting patterns than average industry statistics as pubished by RAA. The reporting pattern is heaviy dependent upon the carrier s case reserving phiosophy (e.g., use of additiona case reserves) and how quicky caims are reserved as permanent tota disabiity cases. 6

7 Company A has been in existence for 15 years and prior to year 10 ony wrote first doar workers compensation coverage in 15 states. Starting in year 10 Company A began to offer high deductibe poicies, a with a deductibe amount of $OO,GOO. The high deductibe premium represents 5 % of tota premium in year 10 and grows to 10% in year 11 and 15% in year 12 and subsequent years. Company A assumes that the high deductibe poicies are a sma percentage of the tota so it does not ater its reserving procedure (which consists mainy of an incurred oss deveopment method based on the historica weighted average deveopment factors). The incurred deveopment projection produces an accurate estimate of reserves for years 9 and prior. As the attached Exhibits -4 dispay, this approach wi substantiay underestimate reserve eves. The reserve underestimation represents over 30% of carried reserves at the end of year 15. This exampe is based on a hypothetica bock of workers compensation nusiness and is intended to highight the importance of separating the high deductibe experience and anayzing it separatey. 7

8 It shoud be noted that in addition to the reporting pattern difference, two other factors wi affect the reserve shortfa using Company A s traditiona approach: 1) The trend for excess osses exceeds the trend in primary osses; and 2) The utimate undiscounted oss ratio for high deductibe poicies generay exceeds the oss ratio for primary poicies. This is argey because investment income wi be substantia for high deductibe poicies. As the above exampe impies, the extended reporting pattern for excess/high deductibe workers compensation products compes the actuary to pace itte weight on the unadjusted traditiona incurred projection method. We woud recommend that the foowing techniques be utiized to estimate reserves for the carrier s high deductibe exposures: 1) Counts times average severity 2); 2) Trended pure premium method*); 3) Expected oss ratio method; 4) Bornhuetter-Ferguson method (B-F); and 5) B-F method - adjusted for off-baance. *) See Funding for Retained Workers Compensation Exposures by Brian Z. Brown and Michae D. Price, CAS Forum, 1994 for a discussion of these methods.

9 Additionay, we woud recommend that medica osses be anayzed separatey from indemnity osses. These two types of osses have different deveopment patterns and much of the excess deveopment in the oder accident years is usuay attributabe to medica osses, Methods 1-4 are widey used and discussed in detai in the actuaria iterature3). We beieve that method 5 is aso used but the particuars of this method are not as we pubished. Therefore, we wi provide a brief description of this method. There are two parameters (assumptions) which are needed to perform B-F cacuations by accident year. A set of d priori oss ratios (which wi vary by accident year based on rate adequacy as we as other factors); and An assumed reporting pattern for incurred osses. When anaysts seect their assumptions, they use their best acnaria1 judgement; however, they wi not know for many years (or possiby not even in their ifetime for excess workers compensation) if these assumptions are correct. Additionay, the assumptions need to be revisited 3, NCCI pubishes data to assist in seecting excess frequency and severity assumptions - see Giam, Retrospective Rating: Excess Loss Factors (PCAS LVIII). Additionay, many carriers can create historica excess experience b;/imposing phantom deductibes on previous first doar caim experience. Methods 1-4 above refer to projections in the excess ayer (i.e., for method 1 the counts and average severity are for the excess ayer). 9

10 annuay, and modified if indicated. The B-F off-baance method incorporates an additiona ste into the traditiona B-F method. This adjustment is documented in Exhibit 5 and invove: comparing actua reported osses to expected reported osses (for a accident years) and adjusting the B priori oss ratios for a portion of the difference in the ratio of actua to expected reportec osses. One potentia shortcoming with the traditiona B-F method is that if actua oss experience is worse (or better) than expected due to an understatement (or overstatement) in the g priori oss ratios, it may take a ong time before this is refected. The B-F adjustment, as dispayed on Exhibit 5, corrects for this phenomenon by adjusting for 50% of the indicated off-baance (i.e., the percentage difference between the actua reported and expected reported osses).4) We seected the 50% for iustrative purposes. We beieve that it is important that the actua oss experience be used (at east partiay) to modify the initia assumptions. In our first exampe on Exhibit 5, we constructed a scenario where the anayst seected an B priori oss ratio of 80%) whereas the actua oss ratio is 100%. We then dispay the corresponding off baance cacuations. For a accident years combined, we woud have expected $1.2 miion of osses to be reported; however, $1.5 miion was actuay reported. This shoud aert the anayst 4, It shoud be noted that anaysis of the data may assist in seecting the off-baance weighting. For exampe, if the ratio of actua to expected osses is ess than one for a accident years, it may impy that the a priori oss ratios are overstated (indicating an offbaance weighting near 1.OO or revision of the a priori oss ratios). However, if there is a trend in the ratio of actua to expected osses it may impy a bias in the reporting pattern (this woud indicate a ow off-baance weighting and a revision to the reporting pattern). In other cases, it may not be cear from anaysis of the data which assumption is biased so a weighting near 50% may not be unreasonabe. 10

11 that one (or both) of the underying B-F assumptions may be incorrect. Underying assumptions shoud be scrutinized, particuary if the ratio of actua to expected osses is either consistenty ess than 1.OO or greater than 1.OO for mutipe accident years. However, it may be difficut or impossibe to determine whether the a priori oss ratios shoud be modified or the reporting pattern shoud be modified. Therefore, we introduce the off-baance cacuation. In the exampe on Exhibit 5, the actua reported osses are 25 % higher than the expected osses. Therefore, we adjust these a priori oss ratios upward by 12.5%, or one haf of the off-baance. We theorize that since actua experience is not consistent with our expectations, either the % priori oss ratios are understated, the reporting pattern is too sow or the experience to date has a reativey arge random eement. We have assumed that 50% of the difference is attributabe to the ;i priori oss ratio assumption. The bottom of Exhibit 5 dispays the revised B-F cacuation and the resutant oss ratio of 93 % for a accident years combined. This adjusted B-F cacuation produces resuts coser to the actua oss ratio of 100% than the initia unmodified B-F cacuation which produces a oss ratio of 86% for a years combined. The accuracy of the off-baance cacuation is dependent upon many factors incuding: The accuracy of the initia assumptions; and The randomness associated with the actua reported osses to date. As mentioned above, we beieve that if the actua osses reported to date are consistenty and significanty different than expectations, then the anayst shoud repeatedy review the assumptions 11

12 underying the B-F cacuation. If the anayst does not have enough additiona information to modify the assumptions, we beieve that the B-F adjusted for off baance shoud be reviewed when seecting utimate oss ratios. We have computed B-F cacuations both with and without an adjustment for off-baance for the foowing scenarios (note that we have assumed that the true oss ratio is 100%): The corresponding cacuation for the adjusted B-F is as foows: Adjusted (for off-baance) B-F Loss Ratio* - AJI Years I Lass Ratio Assumptions Reporting Pattern Less Than Actua Equa to Actua Greater tban Actua 4 Quicker than Actua 72% 77% 83% Equa to Actua 93% 100% 107% Sower than Actua 105% 113% 120% *For 50% of the off-baance II For the exampes we constructed, the adjusted B-F cacuation produces more accurate indications when the expected reporting pattern is accurate. It is aso generay more accurate when the a 12

13 priori oss ratio is understated. As a note, understatement of the ii priori oss ratios is often a concern for reserving actuaries. II. MANAGEnCAREINrTJATVES.. ton of wed Care ntat V The objective of workers compensation managed care can be summed up in one sentence; To combine medica cost containment with optima medica treatment and concurrenty expedite worker re-entry into the work force. The process of managed care has many possibe components, which is why there are many different definitions of managed care foating about. A comprehensive workers compensation managed care program requires committed participation from a interested parties: the insurance carrier or third party administrator (TPA), the medica provider (hospitas, physicians etc.), the case manager, the utiization review vendor, the empoyer and the empoyee. Each participant brings to the tabe a component of the managed care process. For exampe: Insurance carriers and TPA s must be dedicated to proper caims handing. Workoads per examiner shoud be reasonabe (e.g. maximum of ost time fies per caims hander). Caims handing poicies and procedures shoud foster pro-active, investigative, cooperative caims handing that is aways focused on the utimate goa of caim resoution and returning injured workers back to work. Via preferred provider organizations (PPO s) physicians, hospitas, durabe equipment vendors, home heath care providers etc. agree to provide medica goods and services at pre- 13

14 negotiated discounts as ong as one of the providers in the PPO is used. The pre-negotiated discounts are usuay 15% to 25% beow the charges aowed by the egisated workers compensation fee schedue for a given state, if one exists. If there is no fee schedue then the pre-negotiated discounts wi be ess than the usua and customary charges for the area. Discounts typicay vary by type of provider. An orthopaedic surgeon wi often give ess of a discount than an internist simpy due to the aw of suppy and demand. Providers must be focused not ony on proper medica treatment for the injured worker but aso in returning that worker to gainfu empoyment as soon as feasibe (in order to reduce indemnity payments). Thus it is not sufficient to simpy use a typica heath care PPO for workers compensation injuries. Workers compensation PPO s must incude occupationa medicine physicians, providers must be trained on return to work issues, and some types of speciaity physicians, such as obstetricians, may not be necessary at a.. Heath maintenance organizations (HMO s) are aso providers of workers compensation medica services. HMO s provide comprehensive medica care for a negotiated fixed fee per person, payabe per month/year, caed a capitated rate which is paid to physicians for deivery of a heath services to injured workers. The capitated rate is fixed regardess of the amount/type of service rendered. Physicians and other heath professionas are on saary or under contract with the HMO to provide such services at the capitated rate. Injured workers are steered by their empoyer to a primary care physician (gatekeeper) within the HMO who decides upon appropriate medica treatment and refers injured workers to speciaists within the HMO if necessary. 14

15 A case manager is typicay a registered nurse. with a certified case manager (C.C.M.) designation and experience in handing industria disabiity cases. The case manager ensures that proper medica treatment and return to work protocos are appied for a specific type of injury. Such protocos are avaiabe from severa different sources incuding Miiman and Robertson, Interqua and the Commonweath of Massachusetts Department of Industria Accident Study. Many managed care organizations deveop their own interna protocos as we. The case manager deveops a treatment pan for the injured worker based on protoco and the particuar set of circumstances, communicates it to the treating physician, empoyer, empoyee and caims hander and then constanty monitors the treatment process to keep it on track. The case manager wi aso work cosey with the empoyer and perhaps a vocationa rehabiitation speciaist to deveop appropriate ight duty (return to work) programs where necessary. Utiization review is often outsourced to a vendor. The goa here is to infuence, manage, assess, improve and review patient care on an individua case basis. Via utiization review, medica treatment is evauated based upon frequency, duration, and medica reasonabeness and necessity. Utiization review can be conducted on a prospective, concurrent or retrospective basis to pre-certify hospita admissions. The empoyer s roe in managed care is pervasive. Empoyers shoud have we defined ight duty work programs for injured workers incuding a video tapeiibrary of avaiabe jobs, job descriptions with appicabe stated saary and defined duration of job avaiabiity. Empoyers shoud educate empoyees regarding the importance of reporting a injuries immediatey to a supervisor and in expaining the workers compensation system and avaiabe benefits. 15

16 Supervisors must be trained to steer injured workers into the empoyer s PPO or HMO and to immediatey report caims to the carrier or TPA. It is imperative that empoyers maintain effective communication with an injured empoyee via frequent teephone cas, persona visits, cards and incusion in any company sponsored events so that the empoyee knows tbat the empoyer is genuiney interested in their return to good heath. This wi aso tend to keep the empoyee/empoyer reationship from being adversaria, which often eads the empoyee to hire an attorney. Weness programs shoud aso be offered to a empoyees e.g. weight reduction programs, smoking cessation programs and newsettersiteratnre on pertinent heath topics. The empoyee s wiingness to be restored to good heath and gainfu empoyment is critica to the utimate success of a managed care program. The caims examiner, case manager and empoyer must a work together to assure the empoyee that they are receiving the proper medica treatment and that the empoyer is ready for them to return to work the moment they are reeased to do so by their physician. Obviousy the most effective workers compensation managed care program is one where a participants are committed to the common goa of returning the injured worker to fu heath and thus to their job as quicky as possibe. Now that we have described the basic eements of a managed care program, we wi review the resuts of three different studies that measure the savings of different types of programs. We wi then discuss the possibe impact on reserving of different aspects of managed care. 16

17 ,. ost Sav& Due to m V Findings of Actua Studies Presented beow are findings from various studies performed which measure savings generated by using managed care techniques on work reated injuries. rtce Dqartment Workers Compensation Manapedt Pr&. ) The piot project consisted of two programs. The first invoved 17,000 state government empoyees in south Forida. Haf of these empoyees received medica care via an HMO and haf through the traditiona fee-for-service arrangement (known as the contro group) where no managed care initiatives were used. The second program was for 7,500 privatey empoyed workers in the Tampa-St. Petersburg area. Medica care for these workers was provided through a PPO. Loss data for the study consisted of payments on caims with accident dates between June 1.5, 1991 and March 15, Over 5,500 individua caims were incuded. Findings The authors of the study observed significant differences in the acerage costs of injuries treated under managed care versus the traditiona fee-for-service arrangement. In genera the differences were attributed to ower use of hospita services, ower incidence of indemnity caims and fewer and ess costy use of physician services in a managed care environment. ) Forida Managed Care Piot Program; Juy 1, 1994 Fina Report, prepared by Phiip S. Borba, Ph.D., David Appe, Ph.D., and Matthew Fung, Ph.D of Miiman and Robertson, Inc. 17

18 IWO Resuts Average caim costs for the HMO participants were 60% ower than the average caim costs in the contro group. Of this 60% savings: c 6-7 percentage points were attributabe to ower incidence of indemnity caims and shorter duration of indemnity caims * 8-12 percentage points were attributabe to ess frequent use of hospita services k O-5 percentage points were attributabe to fewer days of treatment and fewer numbers of physician treatments 2640 percentage points can be attributed to other aspects of managed care such as payments for medica services were discounted 15% off the Forida fee schedue and HMO participants were treated with a ess costy mix of services. Average caim costs for the PPO participants were 28% ower than the average caim costs in the contro group after area factors were considered. Of the 28% savings: * 7-8 percentage points were attributabe to reduced incidence and duration of indemnity caims c percentage points were attributabe to ess frequent use of hospita treatments 18

19 2-7 percentage points were attributabe to fewer days of service and medica treatments wre Workers Com~ensatronw~~ On Apri 1, 1993, Liberty Mutua Insurance Company and Heathsource New Hampshire became the soe servicing carrier of the New Hampshire assigned risk pan. Heathsource directs the appication of managed care techniques such as negotiated fee reductions with providers, use of ess costy services, recommendations regarding optima treatment patterns and review of invoices for reasonabeness of charges both in regard to amount and appropriateness of procedures in ight of diagnoses. Heathsource has aso introduced weness programs for empoyers. Both Liberty Mutua Insurance Company and Heathsource worked with empoyers to improve their return to work programs. Findings Paid oss ratios after Apri 1, 1993 were 20% to 27% ower than expected based on historica pan experience: b 7 to 12 percentage points of the savings were attributabe to ower than expected average caim costs 6, A Preiminary Evauation of Changes to the New Hampshire Worker s Compensation Assigned Risk Pan as of March 31, 1994 Prepared by Miiin and Robertson, Inc. 19

20 b 14 percentage points of the savings were attributabe to fewer caims and/or more premium than expected. (i.e., the reduction is probaby a resut of oss prevention programs, weness programs and an increase in the premium coected reative to historica eves.)... /NCCI!&&&& for ~cza Itt@tgt of Workers Co- Care Tech&g.& Since 1970, Intracorp has been providing workers compensation rehabiitation and managed care services across the United States and Canada. This study measures the impact of their Eary Assessment workers compensation managed care product which combines eary reporting and intervention with aggressive medica, utiization and return-to-work management by registered nurses using interna protocos. Potentia savings from use of a PPO were not measured. The NCCI studied 38,000 ost time caims in many states from severa of Intracorp s argest customers incuding a muti-state sef-insured empoyer and a state fund. 5,000 of these caims were managed by Intracorp, the others were not. The NCCI measured caim costs from these sources over identica time periods and controed for variabes infuencing caim costs such as state egisation, medica and indemnity infation, empoyee popuation, age and catastrophic caims experience. ) Intracorp/NCCI Methodoogy for Measuring Financia Impact of Workers Compensation Managed Care Techniques. December

21 Findings On average, caim costs dropped about 23 % when case management intervention took pace within three months of accident date. * Managed caims cosed 27% faster than those that were unmanaged * Savings are highest on the ongest, most severe cases and Eary Assessment successfuy seects these cases for management. Whie each study empoyed a different managed care mode and focused on different cost drivers, one item commony measured was the decrease in average caim cost. In ight of the findings of these studies, what woud you say regarding the potentia savings of a managed care program? One question rarey asked is What were the baseine caims handing phiosophies, processes and procedures before managed care techniques were appied? What are we measuring from? If caims handers were simpy bi payers (as does happen sometimes) and a comprehensive managed care mode was introduced to the process then a radica savings coud be achieved. If caims handers are adepty performing their duties and appying certain aspects 21

22 of managed care on their own aready (e.g., trying to propery manage the medica component of a caim) then managed care techniques may have a esser impact on cost. Aso, one eement of the studies to keep in mind is that the evauation periods were not ong enough to capture a medica and indemnity payments on ong-duration caims, which of course are the most expensive workers compensation caims. Even though the various studies dispayed a wide variation in their estimates of managed care savings, a of the programs produced savings of some amount. Thus it appears ikey that impementation of managed care in genera wi reduce future year s oss ratios. This information may be used in seecting a priori ioss ratios for Bornhuetter-Ferguson cacuations when estimating reserve eves. Reserving fmpiicarions As actuaries we must quicky become keeny aware of the cost savings potentia of empoying a comprehensive workers compensation managed care program. Indeed, we wi (if we haven t, aready) be asked by our empoyers and co-workers to measure the savings under a given set of specific circumstances. We say under a given set of specific circumstances rather than in genera because there is no way to accuratey measure the savings in genera. Many questions must be asked before making a measurement. For instance; Is the caims examiner for the carrier or TPA cooperating with the case manager? Does the case manager give the caims examiner appropriate information so that the examiner can set, medica and indemnity case reserves accordingy? Effective communication between the two individuas means more accurate and timey case reserves and increases the chances that the 22

23 injured worker wi be returned to work more quicky. This wi potentiay affect a company s reporting and payment patterns. If a PPO is used: What is the distribution of physicians by type of speciaity? What is the discount by type of physician? What types of physicians are ikey to be visited the most often (e.g. occupationa medicine) and how wi this affect the average physician discount ikey to be achieved? Is there appropriate geographic coverage of the network? (e.g., what is the vaue of having three orthopaedic surgeons in the network, a of them residing in one urban area, if many of your exposures are in outying rura areas at the other end of the state?) What hospita discounts are avaiabe? What is the distribution of medica costs between hospitas and physicians for the types of caims expected to be experienced? In genera, the more comprehensive the PPO arrangement the greater the reduction in utimate osses.. If case management is used to what caims wi it be appied, e.g., a caims incuding medica-ony or a ost time caims or ony catastrophic caims such as spina cord injuries? Wi case management decrease medica costs, on a percentage basis, more for smaer caims (temporary tota and temporary partia) or for arger caims (permanent partia and permanent tota)? If the decrease does vary by injury type then what wi the average decrease be? Wi case management increase or decrease disabiity duration? If the case management process works correcty it is ikey that caims wi be resoved quicker, which impies a speed up in reporting and payment patterns. Aocated oss adjustment expense may be reduced if empoyees are treated such that they do not fee the need to hi an attorney to hep them through the workers compensation maze. Aso, overa medica severities shoud decrease and 23

24 the frequency of medica-ony caims may increase as more injuries are kept from becoming temporary tota. If utiization review is used is there a possibiity of dupicative efforts between the case manager, the caims hander and the utiization review vendor? This may increase the need for IJLAE reserves. How effective is the empoyer at steering injured empoyees into the PPO? Does the empoyer ack a return to work program so that even if managed care enabes empoyees to come back to work more quicky there is no job waiting for them? Return to work programs with ight duty jobs wi reduce utimate costs and the resuting needed reserves. Are empoyees satisfied with the quaity of care they are receiving? Is the empoyee a wiing participant in the process, e.g., do they show up for their medica and rehabiitation appointments? The more they cooperate, the ower utimate costs wi be. How were caims handed in the past? If the insurance carrier or TPA was doing itte in terms of managed care, before they impemented a comprehensive program, the potentia for cost savings is very arge. If they were doing an exceent job of pro-active caims handing prior to managed care then the impact wi be ess. 24

25 The above are ony sampes of questions to be asked. The point is to know the specifics of the managed care mode you are working with and attempt to determine how that particuar mode wi affect reserves in reaity versus how it might ook in theory. II. HA&AGED C&&H&WCIAL ABANGI Contracts with a Managed Care Organization (MCO) can have significant impacts on estimating workers compensation reserves. This section wi briefy describe some MC0 arrangements and their effects on estimating reserves. A) Discounted Fee For Services Discounted fee-for-service refers to a reduction from the providers norma fees for certain groups. Larger groups with significant bargaining power are frequenty abe to reduce medica fees in return for the commitment to channe a arge number of injured workers to a particuar provider. Many companies have been using this type of arrangement with medica providers for severa years. Additionay, in some states, a fee schedue may function ike a discounted fee for service arrangement. This type of arrangement is generay beieved to have a sma impact on tota workers compensation costs, uness impemented with other procedures (i.e., utiization review). Providers may agree to discount services but increase utiization. 25

26 Discounted fee for service arrangements can be refected by modifying a priori oss ratios and expected future deveopment if the anticipated savings can be quantified.@ B) Case Rates Case rates refers to a fat fee per caim for medica costs. Typicay the fat fee varies by type of injury (e.g., ower back sprain). One potentia disadvantage of this method is that it may encourage providers to substitute bed rest as a treatment in pace of other treatments to hea injured workers. Thus, case rates may cause a rise in indemnity costs if not propery managed. As discussed ater, dividend compensation arrangements have been introduced as an attempt to offset this reduction in treatment incentive. Under this ype of arrangement, the savings associated with an MC0 are estimated and a percentage of the savings is paid to the MC0 in the form of a dividend. If the case rates are paid up front, this coud dramaticay speed up the workers compensation medica reporting and payment patterns. Additionay, if case rates are fixed for the ife of the caim, the anayst may consider extracting them from the data and treating them separatey (since future medica deveopment may be minima). If the case rates are ony fixed for 12 months of care after the date of injury (or if case rates are negotiated annuay), standard reserve projection methods may not be as materiay biased. s) Brian Brown and Michae Price in Funding for Retained Workers Compensation Exposures quantified the effect of a future 1% trend reduction for workers compensation medica costs. IBID 2. 26

27 C) Capitated Rates Capita&d rates refers to a fat fee to be charged for a workers compensation caimants for certain or a medica expenses. Capitated rates require significant modification to reserve projection techniques. The extent of the modifications wi vary depending on the extent of capitation. We wi briefy describe the adjustments for various eves of capitation. 1) YCaDitation cm Under this arrangement, the workers compensation carrier pays a fee to an MC0 and the MC0 agrees to provide a medica services (for the ife of the caim) for caims occurring during a certain time period. Under this arrangement, the carrier has in essence transferred% workers compensation medica exposure to the MCO. Therefore, the carriers expected retained unpaid obigation is zero after it has paid the fee (ignoring credit risk and the fact that some caims wi not be covered by the MC0 arrangement). The attached exhibit 6, which is based on a presentation given by Ms. Ruth Bauman of Bue Cross and Bue Shied of Oregon, iustrates the transfer of risk from empoyees to MCO s and finay to physicians under a capitation arrangement. However, in most cases the MC0 wi not be responsibe for: + The ifetime of the caim; 27

28 p A caims (especiay those occurring outside of the state); and p The fu medica expense on catastrophic caims. Therefore the reserving anayst wi need to estimate an accrua for the above items. Under this arrangement the MC0 may be responsibe for: Most medica expenses for a 1 to 3 year period after the injury date of a caim; and The first portion (e.g.. $50,000) of medica costs per caim. In this case the reserving anayst is required to estimate a provision for: Caim payments made after the 1 to 3 year period for a given accident year; and Caim cost above $50,000. Caim payments made 1 to 3 years after the accident date can be estimated based on the company s historica data, if avaiabe. For exampe, caim payments made after 3 years can be compared to payro or premium (both shoud be adjusted to current cost and 28

29 benefit eves). Additionay, an expected amount by caim, or type of caim, can be constructed from the company s historica data. The expected medica payments above a threshod during the first 3 years can aso be computed based on historica caim experience. Historica caims can be projected to utimate vaues as we as to current cost eves, and an average provision by caim (or type of caim) can be estimated. 3) Limited Capitation For this arrangement ony certain types of caim procedures are subject to capitation, and the capitation is ony effective for one year. The procedures outined above for Section C2 - Partia Capitation can be used to estimate reserves. A caim count times average severity methcd aso may be we suited to estimate outstanding reserves after the 1 year capitation arrangement. The severity used in this case shoud be the medica severity for payments in years 2 and subsequent. Additionay, caim counts wi correspond to a caims expected to remain open after the capitation arrangement has ended. D) Dividend Formuas Between Workers Compensation Carriers and the MC0 It appears that many carriers and the MC0 are using dividend pans for the foowing purposes: 29

30 An incentive to the MC0 to return injured workers back to work; b To reward the MC0 for effectivey and efficienty managing care; and b To have the MC0 guarantee payments to carriers if oss experience is adverse. We wi describe two types of dividend programs: 1) An incurred oss ratio pan; and 2) An average severity method. One form of the incurred oss ratio pan invoves comparing the actua reported osses to a target oss provision at intervas 2, 3, and 4 years after the end of an accident year or poicy year. The target oss provision is equa to the actua earned premium mutipied by a target oss ratio (adjusted to refect the estimated percentage of osses expected to be reported at the evauation interva). The dividend is equa to a portion of the amount by which actua osses are beow the target osses. In other words, to the extent that the MC0 is abe to reduce costs, part of the savings wi be shared with the MCO. As a technica note, caim payments above a certain threshod are usuay excuded. Exhibit 7 dispays?. sampe cacuation. This method has severa imitations in measuring savings attributabe to the MCO s invovement, because: 1) The frequency (i.e., the number of caims) is usuay outside the contro of the MCO; and 30

31 2) Caim costs vary depending on the type of injury, and injury type is aso usuay outside the contro of the MCO. Therefore, some dividend pans may deveop expected costs based on an estimated severity (average cost per caim) for the prospective period, rather than in aggregate. The actua number of caims is then mutipied by the severity estimate to determine the target caim costs. This target caim cost can then be compared to the actua reported caim costs to derive the indicated dividend. Exhibit 8 dispays the cacuation for a sampe program based on the average severity method. This average severity pan may resut in the MC0 receiving a dividend even if actua tota caim costs exceed initiay targeted caim costs (cacuated in aggregate based on the number of expected caims). In other words, the greater than expected number of re-ported caims is refected in the target caim costs for this method. This is beieved to be appropriate since caim counts are generay assumed to be outside the contro of the MCO. An additiona modification to the average severity method woud invove computing the target costs based on benchmark average caim costs by type of injury. For exampe, expected average severities coud be computed by injury type (i.e, ICD-9 code combination). For this method, the target costs are computed by mutipying the actua number of caims for each injury type by the expected severity for that injury type. These products are then summed across a injury types to arrive at an aggregate target cost. The actua costs are compared to the target cost to estimate 31

32 the projected savings (and a portion of the savings is returned to the MC0 in the form of a dividend). It is important for the reserving anayst to estimate an accrua for dividends to the MC0 if the anayst s company is using these types of arrangements. IV. -G WOW COP One approach used to estimate capitated rates for workers compensation medica costs which has been deveoped by heath actuaries is to project the workers compensation medica costs for a group of injuries based on heath insurance data. An average cost is then computed based on the probabiity of a certain condition and the associated costs of the treatment for the condition. We wi iustrate this type of anaysis for an industria anke injury. Ihe first step is to anayze the costs for anke injuries in more detai. Possibe combinations of anke injuries inchtde:9) 1) Fractures or Disocations ICD-9 Codes: 823.2, 823.3, 824., 837.0, 837.1, ) Sprain, Sprain-Fracture or Contusion ICD-9 Codes: 845.0, , Heath insurance costs are captured by ICD-9 codes. The ED-9 code refers to the 9th revision of the Internationa Cassification of Diseases. 32

33 3) Laceration ICD-9 Codes: 891.0, 891.1, ) Tendinitis ICD-9 Codes: , , , , ,727.81, ) Traumatic Arthritis, Acute Episodes ICD-9 Code: ) Systemic Disease ICD-9 Codes: Mutipe Mihman & Robertson, Inc. has deveoped Heathcare Management Guideines (HMG) based on data from managed care pans and input from empoyed physicians. These guideines incude ranges of time within which injured workers are expected to return to work by injury type (i.e., grouping of ICD-9 codes). An exampe of these guideines is incuded as Exhibit 9. The guideines aso incude ranges of the duration of care by injury type, as dispayed on Exhibit 10. The Heathcare Management Guideines aso incude frequency and cost statistics for the procedures used in the course of treatment of various injuries. Procedure statistics are deineated by CPT code, which refers to the code assigned to a medica procedure under the Physicians Current Procedura Terminoogy. Exhibit 11 outiis initia care statistics for anke fractures and disocations. As shown on Exhibit 11, it is expected that 80% of a cases wi be initiay treated by an off~cr visit, and 20% wi be treated in the emergency room. The probabiities of various procedures being used for treatment 33

34 are then isted by CPT Code in Coumn (b). Based on these probabiities combined with the expected number of times each procedure wi be required (Coumn (e)) and the expected price per service (Coumn (f)), the expected price for each course of treatment can be derived (i.e., by summing across a CPT codes the product of (e), and (f). The $353 estimated tota cost for initia care is then cacuated (see Exhibit 11) by computing the weighted average cost across both courses of treatment using the treatment probabiities in coumn (a) as weights. The foow-up care for anke fractures and disocations may be treated in three fashions: Competey by primary care physicians; Cosed surgery by a speciaist; and Open surgery by a speciaist. Estimated costs for each of these courses of subsequent treatment are cacuated in the same manner as the initia care cost estimate. These cacuations are outined on Exhibits 12, 13, and 14. Based on optima treatment patterns and the heath insurance data outined above, the foowing costs and treatment probabiities for an anke fracture and disocation are estimated: II Probabiity 1 Course of Subsequent Treatment 1 Cost of Treatment* 71% Therapy by Primary Care Physician $1,280 4% Cosed Therapy by Speciaist 2,900 25% Average Open Surgery by Speciaist 4,900 $2,250 *Incuding the cost of initia care 34

35 It shoud be noted that this is the cost for an optimay managed caseo). It is expected that care wi not aways be optimay managed and some workers wi require more services than expected due to extreme cases. These factors coud be buit into me pricing by adding a oading for additiona costs or procedures (or both). The fina eement which is needed in estimating a capitated rate is the probabiity of a certain type of caim. This may be done through an anaysis of historica caim data (e.g., caim frequency per $100 of payro by injury type). The capitated rate coud then be derived by mutipying the cost of each injury by the estimated probabiity of that injury and cacuating the tota across a types of injuries. V. m OF WQ&J$E&S COMfEMi%TION SYS EEQS From 1983 through 1992 workers compensation countrywide combined ratios ranged from 113 % to 123 %, residua market operating osses soared and severa insurance carriers withdrew from writing vountary coverage. These factors ead thousands of empoyers to opt for sef-insurance to escape workers compensation insurance rate increases and the frustration of being unabe to obtain coverage outside of an assigned risk poo. A system participants procaimed the need for reforms that woud ater the system to truy reduce the cost eves and trends of workers compensation benefits without sacrificing equitabe compensation for the injured worker. Thus was born an era of change. From 1991 through 1995 approximatey 60% to 65% of the states o) The above exampe is based on a presentation by Richard Minifie, ASA, MAAA, of Mihman & Robertson, Inc., tited Deveoping Cap&ion Rates Consistent with Cinica Practice Guideines. 35

36 impemented some type of workers compensation reform, ranging from instituting medica fee schedues to totay overhauing a aspects of the benefit deivery system. Severa other states are currenty deveoping pans for reform. )pes of Reforms Listed beow are exampes of different types of reform and the potentia effect on oss reserves... v. Originay workers compensation benefits were for injuries that arose out of the course of empoyment. Over the years compensabiity has been interpreted more and more iberay by courts, for exampe, considering an injury to be compensabe when it occurs at a softba game after work when the team is made up of empoyees from a common empoyer. Additionay, stress caims have been fied by empoyees due to fear or disike of a feow empoyee and some courts have deemed these to be work reated caims. If a reform can bring compensabiity back into ine with it s origina intent then of course the number of compensabe workers compensation caims shoud decrease. This reduction in frequency shoud reduce future year s oss ratios... &&Q&&y. The duration for temporary tota disabiity can be restricted to fewer weeks, which wi ower indemnity severities. The definition of permanent tota injuries has been narrowed consideraby in some states, e.g., in Forida as of January 1, 1994 tota disabiity is iiited to injuries such as severe paraysis, amputation, major burns or other injuries that woud quaify for Socia Security disabiity benefits. This type of reform may increase indemnity and medica severities for permanent tota it removes the ower 36

37 doar cases from the permanent tota category) whie reducing their frequency. Some states escaate the indemnity portion of tota disabiity benefits by an annua cost of iving factor. Connecticut decided that for injuries occurring on or after Juy 1, 1993 the escaation factor woud no onger appy. This change shoud greaty decrease indemnity severities and shorten the tai on payment patterns.. 4 *.. B. Permanent partia disabiity caims represent the argest share of osses in many states, are among the most compex benefits to deiver, and bring more attorneys into the workers compensation system than any other type of caims. r) These benefits vary greatiy among states and can be based on the degree of impaiint or wage oss or oss of earning capacity. Rather than deve into each type of compensation avaiabe suffice it to say that any major reform deaing with this injury type shoud be studied cosey by reserving actuaries <. Comprehensive reforms often incude these areas. Aternative dispute resoution processes are meant to be a more informa, non-adversaria means to resove caim disputes between empoyers and empoyees without the invovement of attorneys (for either side), i.e., without the need to go to court for a hearing. Other reforms specificay aimed at curbing attorney invovement incude eimination of ump sum awards for caimants (because they are very enticing to paintiffs I) BNA s Worker s Compensation Report, Juy 24, NCCI Report Examines State Differences in Permanent Partia Disabiity Benefits 37

38 attorneys who usuay get one third of the award). Some states have aso imited attorney fees to much ess than one third of the award. For instance, Forida s January 1, 1994 aw imits awards to attorneys for indemnity payments to 20% of the first $.5,ooO in benefits, 15% of the next $5,000 and 10% of the remaining benefits payabe within 10 years and 5% of benefits payabe after that. I*) Obviousy such reforms shoud greaty reduce aocated oss adjustment expense payments as fewer cases wi work their way into the court system.. Medica care cost containment. Various medica cost containment strategies have been impemented in most states incuding empoyer choice of physician, imited provider change, use of medica fee schedues, reguation of hospita charges, mandated utiization and/or bi review and use of other managed care techniques. The Workers Compensation Research Institute has examined the use of such cost containment strategies over the past five years. Exhibit 1.5 shows the types of cost containment measures that were in effect from 1991 to states imited the empoyee s initia provider choice and 40 states paced imits on an empoyee s abiity to change providers. 27 states had medica fee schedues in pace and 22 reguated hospita charges via statute. Ony about 14 states mandated utiization and/or bi review by payers, the workers compensation agency and/or the state fund. I*) BNA s Worker s Compensation Report, November 22, 1993 Lawmakers Approve Reform Package: Aows Managed Care, Limits Attorneys I31 WCRI s Medica Cost Containment in Workers Compensation - A Nationa Inventory