Current location:	Endicott, NY, USA
Current mood:	angry
Current music:	"Funeral/Rebuilding Serentiy", David Newman
Entry tags:	benefits, bills, eobs, free enterprise, free market, health care providers, health costs, health insurance, hmo, insurance, medicare, medicare disability, monolopy power, mvp, mvp health care, remuneration, step response

MVP Health Care: broken

updated: 20070326

as promised, here is a quantitative followup to my earlier blog entry talking about the step response of the medical care system when incrementally burdened with a difficult case, the health management of my wife, Helen. such a technique has been used for characterizing economic systems as well as traditional control settings and has in fact been used to examine medical costs as well as in other medical contexts.

my justification for singling out of MVP Health Care for examination is straightforward. first, i can, since the data are in hand. second, i do not have access to other HMOs performance since they did not care for Helen. third, it's not accurate to say these data are typical of all HMOs, since i do not know that. if MVP were typical, such inferences might be possible. fourth, there is no agreement or condition in the contract with MVP which demands or even requests records of transactions with them to be held in confidence. fifth, MVP no longer provides health care for either Helen or for me.

results

the principal data are presented in graph form. they appear below the cut, since they are large images. methods used to derive these conclusions are recorded in the last section, below the graphs. in the sample, 8 payments were made in full, and 2 more nearly so. these are all home health care supplies or psychiatric interventions, although based upon the sample there are many other cases where payments were not made or made substantially less than requested. 12 requests for payment met with zero responses from MVP. these might be problems of coding on the part of the provider, except that the profile as requested is supposed to be the final disposition of all claims. note that this means MVP is paying here as secondary payer, Medicare having first paid to providers whatever is their allotment. nevertheless, 39 of 61 payments (64%) were less than the full amount requested, and 31 of non-zero payments were less than seven tenths of the original request.

discussion

MVP Health Care's remuneration system to health care providers is either badly broken, or they are playing a cynical game of putting off payments to them as long as possible and, then paying them a fraction of their bill, with no compensation for the cost of doing MVP's imposed paperwork. indeed, if MVP were a small business or in other than a government protected industry, they would be accused by creditors of failing to pay bills. instead, MVP Health Care is using its size and accounting and market powers, approaching those of a commercial monopoly, to browbeat health care providers into dealing with it on its terms, and having them pay for its own mistakes.

either the 12 (of 61) instances of zero payments are in fact MVP refusing to pay legitimate claims, or they are MVP's attempt to "cut their losses" with respect to Helen's patently expensive health care, or they are cases where providers failed to meet MVP's documentation or other standards when submitting claims.

consider each of these possibilities in turn.

first, it's possible MVP's health care remuneration process is simply a huge mess, suffers from inconsistent policy, incomplete training, busted software, or some combination. the trouble is, they are insulated from market forces which would otherwise take down any company comparably inept. this is a "free market" or free enterprise?

second, if MVP is deliberately refusing to pay legitimate claims, they are fraudulent. it is an open question why or why not state and federal agencies don't take a more active role investigating this process. in fact, i think it is time New York's newly elected governor brought some of his authority and values to their doorstep.

third, if MVP is denying claims to providers because Helen's medical care has become "too expensive", they are in violation of their own members agreement and contract, since it has no lifetime ceiling on medical care or costs, apart from whatever subjective constraints are imposed by the mysterious and powerful "utilization committees". perhaps MVP's turnaround on Helen's primary-secondary status was moved by a desire to get themselves a better deal, sticking Medicare with the bulk of Helen's costs. it might not have been that deliberate, but it could be that their discovery and exploitation of the distinction between Medicare and Medicare Disability became important in light of Helen's expenses^*. what has harmed so many is that the costs of recoding and resubmitting the 1600^** or so transactions since MVP declared themselves secondary has been borne by accounting staffs at health care providers, and so must eventually be recovered in their rates, charged back as burden rate to (all) health insurance companies and to patients.

fourth, it's possible claims are submitted with insufficient documentation or standards, resulting in rejections. but then there is a need to explain other related features of the process. for instance, why should seasoned medical professionals and their offices fail to provide adequate documentation in some cases of treatment and not in others, when the procedure and treatment are essentially identical? Helen's care is highly predictable. it's not like she is a candidate for highly experimental surgery or has maladies which demand unusual manipulations. all of her interventions are standard. the trouble is, there are just a lot of them. could it be that MVP is the one arbitrarily varying standard of documentation instance by instance? a health care provider cannot meet standards of documentation they don't know about. and, if they were any other business, they could not control costs as they should if they cannot predict how much effort billing is going to take.

so, in all four instances, MVP Health Care appears to be the cause of this problem. as written earlier, i do not know how widespread this kind of insurance malpractice is in HMOs or elsewhere. to the extent MVP Health Care is a typical health insurer, one would expect to see more of this elsewhere. and that's a big problem. to the extent MVP Health Care is atypical, they should be shut down.

update 20070326: National Public Radio on its Morning Edition carried a story about how doctors' time is being increasing consumed with disputes with insurance companies.

(^*) if the cost of this case did play any part in their flip-flop, they may not have won much, since a good deal went beyond Medicare's allotments and was passed on to them -- all of what's reported here.

(^**) 1600 is quoted rather than 2000 because MVP's announcement of secondary status occurred in late July of 2006. so, given 21 interventions per week, that's about 400 less than if this declaration was made at year's end. of course, the burden on health care providers is understated by this, this the flip-flop caused Medicare to retract its payments, too, because its billing procedures are different when it is primary as it was after MVP's July announcement than as secondary before then.

payments versus provider charges (log-log, showing full payment line)

distribution of amount paid as fraction of request

methodology

a "patient profile" for my wife was obtained from MVP Health Care by telephone request. this covers the period 2005-01-01 through 2006-12-31 which Helen was coinsured by them. her coverage was part of a family insurance plan under which i was also covered. my health care expenses are relatively modest. if readers will recall from the earlier blog entry, Helen also has Medicare and much of the confusion regarding payments was caused by a mistake MVP made in their determination of primary versus secondary status for someone who has Medicare because of a disability. in fact, the profile was requested to have an authoritative source to consult when health care providers approached us with claims regarding payments they were owed. the profile was not sought with the purpose of writing these blog entries nor for purposes of research. in fact, i began the investigation simply wanting to know how much Helen's care cost altogether and how many medical interventions she had.

the "patient profile" is like an "explanation of benefits" report except it covers all transactions with all health care providers who submitted claims. profiles identify the payee provider, date of service, payment amount requested by provider, MVP's fee or payment to the provider, any coinsurance amount we paid, descriptive information, and disposition information along with reason codes. i estimated there were 540 pages of such profile reports, an estimate obtained by measuring height of a subset and then of the entire stack. uncertainty measurements for these are available from the author (by email at jtgalkowski <at> alum <dot> mit <dot> edu) but they gave a worst-case range of 490 pages minimum and 590 pages maximum, estimates obtained by standard non-statistical error propagation.

since tallying and recording all transactions was cumbersome, i resorted to statistical sampling to obtain a representative look at MVP and health care provider behavior. about 10% of the profile pages were chosen at random, sampling without replacement (actually 10%-12%). randomization was introduced by repeated shuffling of the approximately 550 page stack, initially about two dozen times, and then half a dozen times between each election. pages were chosen from an arbitrarily chosen point near the top, the middle, or the bottom, and a conscious but arbitrary choice was made to take another page from the same vicinity (after shuffling) to be sure near neighbor pages were represented. this was repeated 61 times to obtain the observations used and reported.

the results were recorded in Excel (also available from the author). payments by MVP were lessened by any coinsurance paid. no attention was paid to time of payment or order of date-of-service.

$79265 was requested by all health care providers in the sample. MVP actually paid $26280. the mean request is $1299, the median $458. the mean payment is $431, the median payment is $100. a large number of zero responses from MVP to requests skew the distribution.

there are a couple of large requests and one large payment which can be seen in the sample. this is to be expected, as Helen had over a dozen hospitalizations in the two year period. also, none of this reflects any prescription or physician-ordered over-the-counter medications, treatments, or supplies. it does include home health care supplies and visits. i estimate Helen had 2200 separately billable medical and psychiatric interventions in the two year period, an average of 21 per week.