The Accuracy of Dermatology Network Physician Directories Posted by Medicare Advantage Health Plans in an Era of Narrow Networks

Importance  Insurers are increasingly deploying “narrow networks” with fewer contracted physicians both in health plans offered in new state exchanges under the Affordable Care Act and in Medicare Advantage (MA) plans, which are commercial alternatives offered to Medicare beneficiaries. Patients choosing health plans rely on the accuracy of network directories posted by insurers. The MA plans must meet network adequacy requirements, and inaccurate directories of participating physicians might prejudice those determinations.

Objective  To determine the accuracy of MA plan directories of participating dermatologists, and the appointment availability of listed physicians.

Design, Setting, and Participants  Scripted telephone calls were placed to every dermatologist listed in directories for the largest MA plans in 12 US metropolitan areas. The caller sought an appointment on behalf of his fictitious father who had severe itch for several months, asked whether the dermatologist accepted the relevant plan, and asked for the next available appointment date.

Main Outcomes and Measures  Appointment availability and wait time.

Results  Among 4754 total physician listings, 45.5% represented duplicates in the same plan directory. Among the remaining unique listings, 48.9% of physicians were reachable, accepted the listed plan, and offered an appointment for our fictitious patient. Many of the dermatologists listed had incorrect contact information, were deceased, retired, or had moved, were not accepting new patients, did not accept the insurance plan, or were subspecialized. The mean (range) wait time for appointments among the remaining listings was 45.5 (1-414) days. Both the accuracy of network directories and the appointment wait times varied substantially by health plan and metropolitan area. For 1 plan, our caller was unable to obtain an appointment with any listed dermatologist.

Conclusions and Relevance  Medicare Advantage physician directories for dermatology in many areas substantially overestimate the number of in-network physicians available to treat patients with medical skin conditions. These inaccuracies occurred in areas with long appointment wait times and where plans are terminating selected physician contracts. This suggests a lack of capacity that would be exacerbated by further network narrowing. Accurate physician directories are essential for proper oversight of network adequacy, and for patients who rely on these listings to evaluate health plan options during open enrollment.

In the evolving health insurance marketplace, many health plans have increasingly deployed “narrow networks,” reducing the number of contracted physicians or hospitals in a given metropolitan area.¹ Whereas the use of these narrow networks in exchange plans offered under the Affordable Care Act (ACA) might have been expected given pressures to control costs and limit premiums, insurers have taken similar actions in several states to reduce the number of participating physicians in private plans offered as alternatives to Medicare patients known as Medicare Advantage (MA) plans.^2-4 Nearly 16 million of the 54 million Medicare beneficiaries are now covered under private MA plans administered by private insurers,⁵ and physicians in many states have received letters from some insurers notifying them that they are being dropped from MA plans in mid-year without explanation.^6-9

Insurers may have legitimate reasons for limiting the number of physicians in their plans. They point out that narrow networks can help them protect beneficiaries, purchasers, and taxpayers by eliminating physicians with unusually high costs or service use.^8,10 However, the mechanisms used by insurers to evaluate high-cost physicians are imperfect and may not permit them to differentiate between physicians who look expensive because they spend wastefully and those who look expensive because they care for the sickest patients. In addition, some have argued that motives for insurers to restrict physician networks are not necessarily benevolent—they may use narrow networks to make their plans less attractive to enrollees with costly health conditions.^11,12

Federal laws on network adequacy in MA plans require insurers to manage the number, mix, and geographic distribution of contracted physicians to ensure adequate access to covered services. The Centers for Medicare and Medicaid Services has adopted formulas for evaluating network adequacy, specifying minimum provider-to-patient ratios for given specialties that also depend on the number of Medicare beneficiaries, total population, and population density of each county.¹³

In response to reductions in MA physician networks, some stakeholders have sought relief in the courts. A large group of Connecticut physicians sued UnitedHealth Group after being removed from its network without explanation, and they won a preliminary injunction requiring the insurer to give physicians an opportunity to contest their removal through arbitration.¹⁴ A similar lawsuit by the Medical Society of New York is pending in federal court.¹⁵ A bill has also been introduced in Congress that would prohibit plans from dropping physicians in the middle of the year.¹⁶

Many exchange plans being offered under the ACA are also using narrow networks, and similar debates related to network adequacy are playing out in that arena.¹ The ACA and its subsequent regulations stipulate that health plans participating in state and federal exchanges ensure a sufficient choice of providers to their enrollees so that services are accessible without unreasonable delay.

Determinations of network adequacy for specific health plans may depend on the accuracy of the insurer’s list of participating physicians. Patients also rely on those lists of participating physicians, which are typically posted online, to make informed choices when selecting health plans. In recent months, there has been substantial media attention surrounding inaccuracies on these physician lists, especially related to ACA exchange plans. There have been anecdotal reports of physicians appearing on network lists who are not actually contracted participants, retired, or even deceased, as well as physician listings having incorrect specialties or contact information.^17-20

Whereas some inaccuracies in ACA exchange plan directories may be expected given the novelty of those products in 2014 and the last-minute contracting that occurred, the MA marketplace is more mature and should have more stable physician listings. Accurate physician lists are particularly important at a time when MA plans are increasingly deploying narrow networks, making determinations of network adequacy critical for both primary care and many difficult-to-access specialties. Patients considering the purchase of products with narrow networks are dependent on accurate physician lists to determine whether physicians whom they currently use or may need in the future will be accessible in-network.

The specialty of dermatology is relatively small, but it has a well-established physician shortage resulting in long appointment wait times in many geographic areas.^21,22 Dermatologists have been among those physicians receiving termination letters from MA plans without warning or explanation.^23-25 There are concerns that state and federal regulators responsible for ensuring network adequacy and patient access to care might rely on the same lists of in-network physicians that insurers are using to populate their public online directories. Anecdotal evidence and unpublished reports have suggested inaccuracies in those directories, but this has not been documented in larger rigorous studies. We sought to investigate the accuracy of dermatologist listings in directories for the largest MA plans in several metropolitan areas across the United States, as well as to gauge appointment wait times for physicians whose listings are accurate.

The institutional review board of the University of California, San Francisco, reviewed the study protocol and determined it exempt. For this study, we selected the 12 mid-sized to large communities followed in depth by the Center for Studying Health System Change’s Community Tracking Survey, which was initially funded by the Robert Wood Johnson Foundation. These communities represent diverse geographic areas and health care marketplaces across the United States. Metropolitan statistical areas (MSAs) were used to define the boundaries of these communities for this study. Most of the MSAs include multiple counties extending beyond the limits of a central city, more accurately reflecting the medical referral base for a covered population seeking dermatology services (Table 1).

Using MA enrollment data from July 2014, we determined the 3 largest insurers in each of the 12 MSAs by enrollment (Table 1). If an insurer offered multiple plans, those plans were combined to determine total enrollment. We then queried the insurers’ websites to obtain their lists of in-network dermatologists for each of the geographic areas. All queries were performed between July 17 and August 3, 2014. The zip codes of listed addresses were used to determine which physician listings qualified for inclusion by location inside the MSA. If a telephone number was missing, we made reasonable attempts to independently look up the physician’s telephone number on Google and in the American Academy of Dermatology’s directory of dermatologists. Duplicate listings containing the same physician for the same insurer in the same MSA were coded as such.

A scripted telephone call was placed by one of us (A.Q. or M.L.) to each unique physician listing. All calls were placed during varied times of the day and days of the week for each MSA and occurred between August 5 and September 4, 2014. The receptionist answering was told that the caller was seeking an appointment on behalf of his father who had severe itch for several months. The caller also stated that his father had just turned 65 years old and was trying to pick an MA plan. He asked whether the listed physician accepts the relevant MA plan(s) and sees patients with itchy rashes and, if so, when the next available new-patient appointment was. Wait times were defined by the number of calendar days from the date of the call until the date of the appointment offered. Any appointments offered were declined.

If the same physician was listed by 2 or 3 of the insurers, the caller asked about whether he or she accepted each of the plans on the same call. When duplicate listings for a single physician involved multiple office sites, only a single office was called, but if the receptionist accepted the plan(s) and offered an appointment at any other site, that was recorded as an acceptance. Even if the receptionist reported that the physician was in-network only for a subset of that insurer’s MA plans (health maintenance organization vs preferred provider organization), we coded the physician as accepting that insurance plan.

Because we attempted to reach every dermatologist listed by each plan in each area, the results presented do not involve sampling within each plan area, so no confidence intervals or standard errors are reported.

For the largest MA plans in the 12 MSAs, a total of 4811 dermatologist listings were obtained. After removal of nonphysician clinicians (30 physician assistants, 25 nurse practitioners, and 2 podiatrists), there were 4754 physician listings (4408 MDs, 346 DOs) meeting inclusion criteria.

Within health plans, a large number of these listings represented duplicates (2164 [45.5%]) (Table 2). Most of the duplicates were multiple office listings at different addresses for the same physician (1870 [39.3%]). However, there were also many duplicates with the same physician at the same address (294 [6.2%]). Typically, these same-address duplications included slightly different versions of the same name (eg, one with a middle initial, one without).

Among the 2590 remaining unique listings, there were many that our callers were unable to contact (464 [17.9% of unique listings]). Some of these were nonworking or wrong telephone numbers, and others were offices that reported that they had never heard of the listed physician. Several more (221 [8.5% of unique listings]) reported that the listed physician had died, retired, or moved out of the geographic area.

Many of the listed physicians whose offices we reached were not accepting new patients (221 [8.5% of unique listings]). There were also several who were subspecialized and were not willing to make an appointment for a patient with an itchy rash. Many of these listed physicians were dermatologic surgeons (158 [6.1% of unique listings]), but there were also many others (103 [4.0% of unique listings]) who claimed another subspecialty such as pediatric dermatology, dermatopathology, or melanoma.

Fewer than half of listed physicians for each plan (1266 [48.9% of unique listings]) were reached, accepted the listed plan, and offered an appointment for a patient with an itchy rash. Because of the large number of duplicate listings, this translates to only 26.6% of individual directory listings being unique, taking the listed plan, and offering a medical dermatology appointment.

Wait times for the appointments offered by those listed physicians were relatively long (median, 30 days; mean, 45.5 days) (Table 3). The mean wait times varied substantially by geographic area and health plan, from 10.5 days for the 2 dermatologists in Blue Cross Blue Shield’s network in Lansing, Michigan, to 146.0 days for the 2 dermatologists in Excellus Blue Cross Blue Shield’s network in Syracuse, New York. Among larger geographic areas with more physicians, the longest mean wait time was 78.6 days for UnitedHealthcare’s network in Boston, Massachusetts, and the shortest was 17.0 days for Humana’s network in Miami, Florida.

The accuracy of the network directories also varied substantially by plan and geographic area. Among larger MSAs, the UnitedHealthcare network in Boston had especially low accuracy, whereas the Humana network in Indianapolis, Indiana, fared much better. In the less populous MSA of Syracuse, not a single dermatologist listed in the directory of Today’s Options plan by American Progressive would offer an in-network appointment to our caller. Also in Syracuse, of the 12 dermatologist listings by Excellus Blue Cross Blue Shield, only 2 were unique physicians offering in-network general dermatology appointments, and they had a mean wait time of 146.0 days.

Even after removing a large number of duplicate listings, our secret shopper callers were unable to reach and be offered an in-network appointment for an itchy rash with a majority of the dermatologists listed by the top MA plans in a variety of geographic areas across the United States. Some of the listed dermatologists were unreachable because they were retired or deceased, had moved out of the area, or had faulty contact information on the insurer’s website. Many also were not accepting any new patients or had subspecialized practices in dermatologic surgery for skin cancer, pediatric dermatology, melanoma, or dermatopathology. Some offices reported that despite being listed on the health plan’s website, they did not accept the insurer’s MA plans.

The accuracy of physician directories varied by geographic area and health plan. For some less populous metropolitan areas, such as Lansing and Syracuse, the combination of a small number of in-network dermatologists and inaccurate physician directories leaves patients with few if any choices when they need a dermatologist for a medical skin condition.

Our findings also confirm that wait times for dermatology appointments remain long in many geographic areas. There is no accepted, evidence-based metric for an appropriate wait time to see a dermatologist, although patients and their referring physicians would certainly have different expectations for a routine skin check, a symptomatic, itchy rash, and a lesion suspicious for skin cancer. We do not believe that the observed appointment wait times of substantially more than 1 month in many areas for our fictitious patient with severe, widespread itch would be acceptable to most people.

Our findings of inaccurate physician directories and long dermatology appointment wait times in many areas come at a time when many health plans are narrowing their MA networks by reducing the number of in-network physicians. There has been concern voiced about the proprietary methods that health plans are using to select physicians for termination and whether those methods themselves may limit access for patients with more costly health conditions. Our study cannot address those specific concerns, but our data do support the hypothesis that the large physician directories posted by many health plans exaggerate perceived access to dermatologists for MA patients. Furthermore, the long in-network appointment wait times observed for many of these plans suggest a lack of capacity that can only be exacerbated by further network narrowing.

Our study has several limitations. We do not know whether all of the physicians listed in these directories as dermatologists were trained in dermatology residencies or board certified. We cannot assess whether the observed appointment wait times would have been shorter if our fictitious patient had had his primary care provider call on his behalf to request a more immediate appointment. We are unable to assess whether the findings in these 12 MSAs are generalizable to the entire United States, but these MSAs do include a diverse slice of the country and a mix of many small and large health plans. We are also unable to assess whether MA physician directories for other specialties have similar error rates, but we have no reason to believe that contracted dermatologists are more difficult to track and list than others. It is possible that some of the offices that reported not accepting a particular plan were mistaken and unaware of their contracted status, but even if so, real patients seeking appointments would face the same misinformation that our callers faced.

We have not been able to obtain the reports that health plans submit to the federal government to establish their network adequacy, so we are unable to assess whether the large proportion of duplicate listings in the directories are misleading only to patients viewing the website or whether regulators are also misinformed by the duplicate listings. It is possible that the government uses unique identifiers (such as Medicare unique physician identification numbers) to remove duplicate listings for their own analyses. Even with the duplicate entries removed, a majority of listings still were not available to our fictitious caller seeking an appointment.

In addition to directory errors, our observed long wait times for health plans whose networks may be declared adequate may relate to the fact that network adequacy measures simply count dermatologists and do not account for full-time equivalents. Many of the areas with long wait times have 1 or several academic medical centers, where many dermatologists who primarily do research or teaching may only see patients infrequently.

This study did not assess the accuracy of physician directories for health plans being sold on the state and federal exchanges as part of the Affordable Care Act. There have been anecdotal reports of similar problems with their physician directories, and some hypothesize that they are even less accurate than MA directories because of the newness of exchange plans rolling out this year. Insurers have argued that it has been difficult for them to maintain accurate directories because of last-minute physician contracting as these new plans rolled out,^19,26 and they have even blamed physicians for failing to proactively provide updated information.^19,27

California regulators are investigating Anthem Blue Cross, as well as Blue Shield of California, for inaccurate physician lists related to ACA exchange plans,²⁸ and the California legislature recently passed a statute increasing state oversight of network adequacy.²⁹ Because many exchange plans also incorporate narrow physician networks and because a regulatory framework to establish network adequacy for those plans is emerging, the accuracy of their physician directories is similarly of clear importance.

Our findings confirm that MA physician directories for dermatology in many areas substantially overestimate the number of in-network physicians available to treat patients with medical skin conditions. These inaccuracies occurred in areas with long appointment wait times and where physician contracts with MA plans are being terminated without cause as part of network narrowing. Because patients rely on these network listings to evaluate health plan options and because policy makers may use them to determine network adequacy, health plans should make these directories accurate and keep them up to date. Improvements could include cleaning up inaccurate information, removing inactive listings, identifying duplicates clearly, and collecting and displaying information on the number of patient care hours, acceptance of new patients, and subspecialty status.

Accepted for Publication: September 19, 2014.

Corresponding Author: Jack S. Resneck Jr, MD, University of California San Francisco Dermatology, Box 0316, San Francisco, CA 94143-0316 (resneckj@derm.ucsf.edu).

Published Online: October 29, 2014. doi:10.1001/jamadermatol.2014.3902.

Author Contributions: Dr Resneck had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Resneck, Brewster.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Resneck.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Resneck.

Administrative, technical, or material support: Brewster.

Conflict of Interest Disclosures: Dr Resneck serves on the Board of Directors of the American Academy of Dermatology and the Board of Trustees of the American Medical Association. No other disclosures are reported.

Disclaimer: The views expressed in this article are those of the authors and do not necessarily represent the views of the American Academy of Dermatology, the American Academy of Dermatology Association, or the American Medical Association. Neither Mr Brewster nor any other American Academy of Dermatology Association staff were involved in performing survey calls or data analysis.