Health Occupations Outlook

Transcription

1 Health Occupations Outlook Prepared for the Louisiana Health Works Commission 2012

2

3 2012 Health Occupations Outlook Table of Contents Executive Summary 1. Introduction 2. General Background a. Size and Scope of the Health Care Sector in the U.S. b. Patient Protection and Affordable Care Act i. Individual Mandate ii. Requirement to Offer Coverage iii. Expansion of Medicaid iv. Health Insurance Exchanges v. Pre-existing conditions vi. Incentives for Primary Care c. Review of Relevant Literature i. Private Insurance Crowd Out ii. Moral Hazard and Health Insurance iii. Impact of Adverse Selection on Demand for Care 3. Primary Care Physicians a. Background b. Measuring Employment i. Occupational Employment Statistics ii. Licensing Data iii. Number of Primary Care Physicians iv. Physician Hours c. Key Changes in the Health Care Industry i. Medicare ii. Medicaid iii. Accountable Care Organizations iv. Patient-Centered Medical Home d. Review of Relevant Literature i. Elasticity of Physician Labor Supply ii. Physician Participation in Medicaid Programs 4. Clinical Laboratory Technologists and Technicians a. Background b. Measuring Employment c. Effect of Minimum Education Requirements 5. Medical Coders a. Background b. Measuring Employment c. I nternational Classification of Diseases 10th Revision d. Electronic Billing Systems and Electronic Medical Records 6. Health Works Surveys a. Overview i. Employer Survey ii. Physician Survey b. Primary Care Physician Results i. Physician Inventory ii. Physician Payment Types iii. Changes in Payment Type iv. Physician Training and Recruitment v. Open-Ended Physician Responses c. Technologists and Technicians Results Division of Economic Development i

4 Table of Contents Table of Contents i. Technologist and Technician Inventory ii. Technologist and Technician Education d. Medical Coders Results i. Medical Coder Inventory ii. ICD-10 e. General Results i. Patient Capacity ii. Changes in the Healthcare Industry 7. Employment Forecast a. Primary Care Physician Forecast b. Clinical Laboratory Personnel Forecast c. Medical Coders Forecast 8. Conclusions 9. Works Cited Appendix A: Employer Survey Instrument Appendix B: Physician Survey Instrument ii LSU E. J. Ourso College of Business Administration

5 2012 Health Occupations Outlook Executive Summary This study combines existing research with original surveys of Louisiana health care employers and Louisiana physicians to assess the impact of upcoming demographic and policy changes to the health care industry on the health care workforce. The current workforce is examined and a forecast is developed to show how that workforce will change between now and In particular, this study considers the supply of and demand for labor for three occupation groups: primary care physicians, clinical laboratory personnel, and medical coders. Key findings for primary care physicians: jj Traditional employment statistics significantly underestimate the number of physicians j j The total number of primary care physicians will grow from 5,923 in 2012 to 7,396 in 2020, implying a need to fill 1,473 new positions over the 8-year period jj An additional 1,419 primary care physicians will be needed to replace physicians who will retire or otherwise stop practicing in the state bringing the total demand for primary care physicians over the 8-year forecast horizon to 2,892 jj This level of demand implies an average annual demand for 362 new physicians in these primary care practice areas greater than the total number of physicians across all specialties graduating from Louisiana s medical schools on an annual basis Key findings for clinical laboratory personnel: jj The total number of clinical laboratory technologists will grow from 2,870 in 2011 to 3,303 in 2020, implying a need to fill 433 new positions over the 9-year period j j An additional 472 new technologists will be needed to replace existing workers bringing total demand for technologists over the 9-year forecast horizon to 905 jj The total number of clinical laboratory technicians, will grow from 2,320 in 2011 to 2,776 in 2020, which will add 456 new positions over the 9-year period jj An additional 380 new technicians will be needed to replace existing workers bringing total demand for technicians over the 9-year forecast horizon to 836 jj Responses from employers balancing the tradeoff between cost and quality indicate that no industry setting would prefer to require all technologists to have a bachelor s degree or all technicians to have an associate s degree Key findings for medical coders: jj The total number of coders will grow from 2,360 to 3,116 between 2011 and 2020, implying a need to fill 756 new positions over the 9-year period j j Another 482 coders will be needed to replace existing workers bringing the total demand for coders over the 9-year forecast horizon to 1,238 jj While the cost and time required to prepare for ICD-10 will surely be a hurdle, employers do not expect the transition to ICD-10 to significantly reduce the number of existing workers. Division of Economic Development 1

6 Introduction/General Background 1. Introduction As in other industries, providing high quality services in the health care industry requires the right combination of workers in the right places at the right time. However, many occupations are unique to the health care industry and require significant training, creating a more acute need for public and private efforts to anticipate the future need for workers and develop plans to develop the appropriate workforce. The health care industry in the United States is changing rapidly, including the aging of the Baby Boomer generation and the recent federal health reform legislation, which will extend coverage to millions of Americans who were previously uninsured. These changes present new challenges in efforts to anticipate the future workforce needs of the health care industry. This study reviews major changes to the health care industry with an emphasis on understanding how supply and demand for several occupation groups may change in the future. Using insights from existing literature as well as results from two original surveys, the 2012 Health Occupations Outlook quantifies and characterizes the Louisiana supply and demand of labor for primary care physicians, clinical laboratory personnel, and medical coders. The most important inputs required for health care provision are the health care workers themselves, and physicians in particular play a central role in that care. Physicians are required to hold either a Doctorate of Medicine, or the less common Doctorate of Osteopathy. Both of these degrees require a significant investment of time and money to obtain, and Congress limits the number of residency slots funded by Medicare a key source of support for these programs. An argument can be made that restrictions on residency slots combined with less interest among medical school graduates in becoming primary care physicians has constrained the supply of primary care physicians. The problem is particularly acute in rural areas, as many physicians prefer to locate in urban areas. Because of demographic changes and the anticipated shift toward a greater reliance on primary care as a cost control mechanism, there is concern that the supply of primary care physicians may fall far short of demand in the future. In addition to physicians, the health care sector employs many other types of workers performing a wide variety of critical tasks. Medical laboratories employ staff to analyze samples taken from patients, with the goal of assisting physicians in making diagnoses. These laboratory personnel require less training than physicians and have much lower salaries. The Bureau of Labor Statistics classifies all clinical laboratory personnel into clinical laboratory technologists and clinical laboratory technicians, with the distinction being that technologists generally have more training and perform more complex tasks. This group of occupations was selected for inclusion in the study because of the potential policy changes being considered in Louisiana that would require more education for some of these workers. A final group of workers facing changes unique enough to warrant inclusion in this year s report are medical coders. This allied health profession has traditionally been responsible for determining the codes for diagnoses and treatments that are used by insurance companies, Medicare, and other health payers to determine reimbursements for medical services. The current coding system is referred to as the International Classification of Diseases, 9th Revision (ICD-9), but the U.S. Department of Health and Human Services has set a date of October 1, 2013 for most health organizations to begin using the new ICD-10. This new system is substantially more detailed and will require medical coders to have knowledge of anatomy, physiology, and medical terminology, which many do not currently have. In response to these specific changes in Louisiana s health care industry, the 2012 Health Occupations Outlook was envisioned by the Louisiana Health Works Commission. The study includes an original survey of primary care physicians in Louisiana as well as health employers in the state. Questions included in the survey of physicians focused on the supply of primary care physicians in the state, while questions on the employer survey focused on employer demands for primary care physicians, clinical laboratory personnel, and medical coders. This report first reviews important background information on the health care industry and the changes currently affecting it, and then summarizes findings from academic research relevant to anticipating the impact of major changes to the industry. Using survey results and existing studies, this research provides a forecast of the supply of primary care physicians, clinical laboratory personnel, and medical coders in the state. In addition, results from the employer survey are combined with Louisiana s Occupational Forecast to forecast demand for each occupation group. 2. General Background While all Americans have some level of first-hand experience with the health care system, the high level of complexity created by public and private insurance alternatives and the continuing advances in the provision of care prevent most from having a complete understanding of it. While a complete discussion of the health care system in the United States is beyond the scope of this report, this section provides information about characteristics of the health care market that are most relevant to this study. Health and health care issues have dominated the public policy arena for at least the last four decades, with numerous proposals at both the state and federal level to change the way health care is delivered and paid for most recently, with the passage of the Patient Protection and Affordable Care 2 LSU E. J. Ourso College of Business Administration

7 2012 Health Occupations Outlook Figure 2.1: National Health Care Expenditures, Source: Centers for Medicare & Medicaid Services, National Health Expenditure Data and Projections Act, signed into law on March 23, The ACA introduces a number of fundamental changes in how the health care industry functions. A primary goal of the ACA is to expand access to care. However, partly motivated by the historically rapid growth of this sector of the economy, this legislation also represents an attempt to bend the cost curve downward. We will discuss some of the salient features of the health care sector and the Affordable Care Act in sections below. Size and Scope of the Health Care Sector in the U.S. Both the size and scope of the health sector of the U.S. economy has grown rapidly over the last several decades. Since 1970 spending on health care has risen from just a little under $75 billion to almost $2.6 trillion in This means that current spending on health care is about 35 times what it was in 1960 (Centers for Medicare and Medicaid Services, 2012). Further, the most recent estimates from the Centers for Medicare and Medicaid Services (CMS) project that spending will rise to approximately $4.8 trillion by the year 2021, 64 times what it was in 1960 (Figure 2.1). These projections by CMS include the estimated effects of the Affordable Care Act. In 1960, individuals and families devoted only about $1 out of every $20 to health care. By 1980 health care consumption had grown to the point where $1 out of every $11 was allocated to health care. Currently, about $1 out of every $6 spent on goods and services in the U.S. is spent within health care sector. This can be seen in Figure 2.2, which depicts the rapid growth of health care spending as a fraction of the total value of all goods and services produced in the economy (GDP). In 1970, health care spending represented about 7 percent of GDP. Relative spending on health care increased steadily during the 1970s and 1980s, but leveled off during the 1990s and once again between 2004 and According to CMS, health care spending in the U.S. grew about 3.9 percent in 2010 following the record slow growth of 3.8 percent in By 2010 total health care spending amounted to approximately 17.9 percent of GDP. Projections for future spending suggest that health care will continue the trend of continuing to represent a larger fraction of total spending in future years. Interestingly, the current 2021 forecast exhibits slower growth than projections from before the passage of the Affordable Care Act such as the forecast, which was the last forecast released prior to the passage of the Act. Division of Economic Development 3

8 General Background Figure 2.2: National Health Care Expenditures as a Fraction of GDP Source: Centers for Medicare & Medicaid Services, National Health Expenditure Data and Projections Why did health care expenditures as a fraction of GDP level off during the 1990s? Some claim that it resulted from the movement of a large portion of the population away from the fee-for-service (FFS) system toward more intensive reliance on managed care e.g. HMOs (Folland, Goodman, & Stano, 2007). Traditional FFS systems do not provide patients with incentives to shop around for less expensive care and they do not encourage cost competition among providers. By selective contracting with hospitals and physicians, combined with more stringent utilization controls, HMOs provided less expensive health care by developing a new combination of inputs to providing care. But, the success of managed care was temporary it merely caused a one-time decrease in expenditures due to the shift from one combination of inputs to another. However, time has shown that their costs tend to grow at the same rate as FFS costs. Consequently, the growth in health care costs resumed. Consumer demand also limited the cost-saving benefit of HMOs as they became dissatisfied with cost rationing constraints imposed by typical HMOs and moved toward policies with less stringent gate-keeping policies. Another way of measuring how important the health care sector has become is to examine per capita spending. Figure 2.3 shows, as expected, a fairly steady increase in the amount spent per person on health care. In 1970, the $75 billion Americans spent on health care translates to about $365 per person. By 1980 per capita spending had increased to about $1,126 per person; increasing rapidly to a little over $8,000 in Figure 2.4 reveals that health care spending in Louisiana has generally mimicked the trends at the national level. Between 1991 and 2009, spending on health care in the state rose from about $11.3 billion to approximately $30.5 billion 2.7 times what it was in At the national level, over this same time period, spending increased from $791.5 billion to $2.5 trillion roughly 3.2 times what it was in Clearly, the health care industry represents a large and significantly expanding component of the U.S. economy. Explanations for this rapid expansion of the health care industry are legend, but by expressing health care spending as a percent of GDP, we net out the effects of overall inflation, leaving us with three broad categories of explanations for the secular increase in share of national income devoted to health care. First, steady increase in expenditures may be the result of an equally steady increase in the real demand for health care 4 LSU E. J. Ourso College of Business Administration

9 2012 Health Occupations Outlook Figure 2.3: National Per Capita Health Care Expenditures, Source: Centers for Medicare & Medicaid Services, National Health Expenditure Data an increased utilization of health care services such as, for example, more frequent visits with providers, more exams and more elective procedures. It is certainly plausible to assume that, over this period of time, individuals have become more concerned and conscientious about their own health, leading to an increase in the demand for medical care. So long as medical care is a normal input in the production of health, increases in demand for health will cause the demand for medical care to rise. To the extent that health care in general is considered a normal good, it is not surprising that spending on health care rises with real income (a positive income elasticity of demand). But, since aggregate health care spending has been increasing faster than national income, health would have to be among a class of goods economists refer to as luxury goods, which simply means any good or service for which the percentage increase in spending exceeds the percentage increase in income. Thus, if health care falls within the class of goods for which the income elasticity exceeds 1, the historical trend in health care spending is not a matter of grave concern it merely reflects choices by families to devote a larger and larger fraction of their income to health. Is there any evidence about whether health is a luxury good? While economic studies are mixed on this issue, one study using standardized time series data on health care spending in OECD countries (including the U.S.) found that the long run income elasticity of demand for medical care exceeds unity, which means at the margin, health care is a luxury good. The second possibility is that individuals may be demanding higher-quality health care services. Additionally, many of these higher quality services were not available in the past because they are the result of recent and recurring technological advances for example, MRI machines, laser surgery, robotic surgery, organ transplants, new generation drugs for heart diseases and new treatments for numerous other diseases. It is clear that technological advances in medicine have raised the quality of health care since the 1960s and higher quality generally commands a higher price. Economic theory and empirical analyses suggest individuals are generally willing to pay more for higher quality. Finally, health care inflation may be persistently higher than the overall inflation rate. Reasons offered for the tendency of health care inflation to exceed overall inflation focus on the effects of rising demand associated with higher real incomes, the increased prevalence of third-party payers (insurance) and the substantial increases in demand caused by the long run Division of Economic Development 5

10 General Background Figure 2.4: National and Louisiana Health Care Expenditures, Source: Centers for Medicare & Medicaid Services, National Health Expenditure Data expansion of government funded programs like Medicaid and Medicare. Medical care prices have historically grown faster than prices overall. Figure 2.6 shows the pattern of medical care inflation compared with overall inflation. 1 Relative to the overall CPI, medical care inflation started to diverge from the overall CPI in the mid-1980s. By 2010 the medical CPI had risen to compared to for the overall CPI. For example, between 2009 and 2010, the medical care CPI increased by 3.4 percent while the overall CPI increased by merely 1.7 percent. It may be important to point out that indexes of inflation suffer from an inability to fully capture the effects of quality changes. To the extent that quality has been rising, the inflation index will overstate the true rate of inflation. Relative to all goods and services measured by the CPI, technological advances in the medical sector have been truly impressive. Thus, measured changes in the medical CPI may significantly overstate the true level of inflation occurring in that sector of the economy. The persistently high level of inflation for medical services is a common problem for maintaining health care programs. Over the last few decades the rising costs of providing medical care services has led to the implementation of numerous costcontainment initiatives by government, both state and federal. It is no surprise that per capita spending in Louisiana also rose steadily over this same period (Figure 2.5). However, while total spending on health care in Louisiana rose at a somewhat slower rate than overall spending at the national level, per capita spending in the state increased at virtually the same rate as it did nationally. In 1991, health care spending per person in the state was $2,646 compared to $3,139 nationally. By 2009 per capita spending in Louisiana had risen to $6,792 compared to $8,136 nationally 2.6 times what it was in 1991 for both the state and the nation. An additional artifact of the health care industry that is important for understanding the market for primary care physicians is the sources of payments. Because Medicare and Medicaid have historically paid lower fees to physicians relative to private insurance for performing the same services, physicians often attempt to limit the number of patients covered by public insurance in favor of serving patients with 1 The medical care CPI includes prescription drugs and medical supplies, physicians services, eyeglasses and eye care, and hospital services. 6 LSU E. J. Ourso College of Business Administration

11 2012 Health Occupations Outlook Figure 2.5: Louisiana Per Capita Health Care Expenditures, Source: Centers for Medicare & Medicaid Services, National Health Expenditure Data Note: White-shaded area denotes projections ( ) private insurance. Figure 2.7 shows the historical trends of Medicare and Medicaid gradually playing a larger role in the overall market for health care. If doctors are currently near their limit for accepting Medicaid patients, newly enrolled individuals may still face limited access to care unless payment changes are big enough to entice providers to accept those patients. Patient Protection and Affordable Care Act The ACA represents the most comprehensive change to the health care system since the creation of Medicare and Medicaid in the mid-1960s and as originally drafted would extend coverage to 32 million more Americans by 2016 (Congressional Budget Office, 2011). The provisions of the law are intended to expand health care coverage, control health care costs and improve health care delivery. While ACA contains literally hundreds of detailed provisions, in this section we summarize the more salient features of the law. 2 Individual Mandate. The individual mandate requires all U.S. citizens and legal residents to purchase health insurance. 3 The size of the mandate increases between 2014 and 2016 and when fully phased in will reach the greater of $695 or 2.5 percent of annual income per individuals, but is capped at $2,085 per family. Part of the justification for imposing the individual mandate is to spread the costs of care for the very sick over a larger group of individuals. The mandate is also an essential element for ensuring the viability of insurance given that insurance companies are prohibited under the ACA from denying coverage to individuals with pre-existing conditions, which would encourage individuals not to purchase insurance until after being diagnosed with a condition that is costly to treat. While a number of states filed suit challenging the constitutionality of the individual mandate, the U.S. Supreme Court upheld its constitutionality thereby maintaining it as an essential part of the ACA. Requirement to Offer Coverage. The ACA also imposes on firms with 50 or more employees a requirement to offer health insurance coverage for their employees. Penalties will be levied against employers that do not offer coverage, effective January 2 For a far more detailed summary of the ACA see The Henry J. Kaiser Family Foundation (April 15, 2011). 3 Some people will have an exemption from this mandate, including members of Native American tribes, members of religions who are opposed to modern medicine (e.g. Amish), people who are incarcerated, and undocumented immigrants (Kaiser Family Foundation, 2012c). Division of Economic Development 7

12 General Background Figure 2.6: CPI and Medical Care CPI, Source: Bureau of Labor Statistics (2012a) 1, Employers with less than 50 full-time employees will be exempt from such penalties. Further, for all employers with more than 200 employees, the ACA requires them to automatically enroll their workers in any health insurance plans offered by the employer. Employees may not opt out of that coverage. Expansion of Medicaid. The ACA mandates expansion of Medicaid for all states raising the threshold for eligibility to those with incomes up to 133 percent of the federal poverty line. Currently, states have the authority to set the threshold for Medicaid eligibility, and Louisiana s threshold is set at 13 percent of the federal poverty level. The original act included a severe penalty for states that failed to participate, with a noncompliant state losing the federal share of current Medicaid program expenditures. However, the U.S. Supreme Court concluded that such a severe penalty would be economic dragooning and ruled that it would be unconstitutional to withhold federal funding for existing state Medicaid programs as a consequence of not expanding the program (National Federation of Independent Business et al. v. Sebelius, Secretary of Health and Human Services, et al., 2012). The increased costs associated with this significant expansion of Medicaid are largely supported by funding appropriated by the health reform legislation. States will receive 100 percent federal funding for 2014 through 2016, 95 percent in 2017, 94 percent in 2018, 93 percent in 2019 and 90 percent federal funding thereafter. This cost structure makes state participation in the expansion very attractive from a budgetary perspective. However, it remains unclear if Louisiana will expand coverage under the plan. John Holahan and Irene Headen with the Urban Institute estimate what the impact of ACA will be on new Medicaid enrollees for each state (Kaiser Commission on Medicaid and the Uninsured, 2011). They use the 2007 and 2008 Current Population Survey (CPS) to estimate the new Medicaid enrollee impact for Louisiana under two participation scenarios. The authors first assume moderate levels of participation (57 percent) among the newly eligible uninsured and lower participation for other coverage groups. Assuming a moderate participation rate, the authors estimate that Louisiana will experience an increase of 366,318 new Medicaid enrollees as a result of ACA s expansion of Medicaid to all individuals under age 65 with incomes up to 133 percent of the federal poverty line. In an enhanced outreach scenario (75 percent participation rate among the newly eligible that are currently uninsured), 8 LSU E. J. Ourso College of Business Administration

13 2012 Health Occupations Outlook Figure 2.7: Type and Percentage of Total Health Expenditures, Source: Centers for Medicare & Medicaid Services, National Health Expenditure Data Louisiana can expect to see 507,952 new Medicaid enrollees. However, this published estimate potentially overstates the prevalence of uninsured adults who would be covered by a Medicaid expansion. According to the Current Population Survey, only 713,000 non-elderly adults fall below 135 percent of the federal poverty line (Census Bureau, 2010). More recent data from the 2011 Louisiana Health Insurance Survey indicate that roughly a third of those individuals are already covered by public insurance, just under a third covered by private insurance and 38.5 percent are uninsured. It is true that some newly eligible individuals who currently hold private insurance may opt for free coverage through Medicaid. With less than 300,000 uninsured adults newly eligible, and less than 250,000 covered by private insurance, nearly all of the privately insured would have to switch to Medicaid to reach the upper end of previously published estimates. A more moderate estimate closer to 350,000 is probably much more plausible even after accounting for the crowding out of private insurance. Even with that more moderate estimate, an expansion of Medicaid following the ACA guidelines would significantly increase access, but raise capacity concerns for access to care in Louisiana. Health Insurance Exchanges. The ACA promotes the creation of state-based, standardized health care plans offered through health insurance exchanges, which are intended to facilitate the expansion of health insurance coverage to more individuals and families. Exchanges will be administered by governmental agencies or non-profit organizations through which individuals and small businesses (with less than 100 employees) can purchase health insurance. Exchanges will not be insurers and will not bear the risk themselves. Instead these exchanges will determine which insurance companies will be allowed to participate. There will be five major requirements for insurance companies participating in an exchange as spelled out in the ACA. First, guaranteed issue will prohibit insurance companies from refusing to insure any individual. Second, limits will be set on price variations prices will vary according to a set of four factors, but will not vary by more than a factor of 10. Third, plans issued by insurance companies participating in an exchange will be offered in four-tiers ranging from bronze to platinum with no limits on out-of-pocket expenses. Fourth, strict regulations will be set on rescission based on pre-existing conditions. Finally, insurance companies cannot establish lifetime or annual dollar limits on coverage. Insurance companies have been willing to accept these Division of Economic Development 9

14 General Background constraints because of the Act s inclusion of the individual mandate. The individual mandate will permit insurers to spread the financial risk of all new insured individuals with pre-existing conditions among a substantially larger pool of newly insured low-risk individuals. Thus, at least theoretically, the net risk to insurance companies will not increase as a result of the requirement to insure all individuals. The exchanges also mitigate the impact of the individual mandate by offering a subsidized purchase for lower income individuals. The ACA provides subsidies to those with incomes greater than 133 percent but less than 400 percent of the poverty line who buy health insurance through a health insurance exchange. At present, most states have begun the process of setting up their own health exchanges. Louisiana is one of a small group of states not creating a health insurance exchange (The Advocate, 2012). If Louisiana ultimately decides against creating its own exchange, the ACA provides for a federallyadministered exchange for Louisiana residents and businesses. Pre-existing Conditions. As mentioned previously, the ACA prohibits health insurance companies from denying coverage for pre-existing conditions. Currently, health insurance companies can deny insurance coverage for medical conditions that existed prior to issuance of the insurance policy, such as cancer. Denial of insurance for pre-existing conditions helps insurance companies manage the adverse selection problem associated with individuals who apply for insurance only when they need health care. However, given an individual mandate to have health insurance, the incentive for people to game the system in such a manner is reduced, since there is an added cost to remaining uninsured. increased to Medicare levels in 2013 and But, similar to the doc fix issue, it may be politically difficult for Congress to cut payments to physicians in 2015 (Robert Wood Johnson Foundation, 2010). If this change becomes permanent, it may lead to an increase in the number of physicians willing to offer primary care coverage to the Medicaid population. It is anticipated that the impact of this specific rule will be relatively small in Louisiana, since the state s Medicaid reimbursement levels are currently about 90 percent of Medicare rates (Zucherman, Williams, & Stockley, 2009). 4 The Act also includes a 10 percent increase to Medicare fees for primary care physicians in the years 2011 through 2016, which might partially address the primary care physician shortage. Also, the ACA includes monetary incentives for medical students to select primary care as their practice area. These incentives include scholarships, loan repayment programs combined with the student s agreement to practice at least initially in an underserved area. Review of Relevant Literature Private Insurance Crowd Out. To better understand the potential implications of a large expansion of Medicaid, a brief discussion of private insurance crowd out is warranted. Crowdout occurs when privately insured individuals drop their policies in favor of public insurance when they become eligible for Medicaid. A recent study Gruber and Simon (2008) summarizes literature on the topic from the past decade with the following table highlighting key findings: Pre-existing conditions exclusions are also relevant to the way health insurance affects labor market mobility between jobs. Denial of coverage due to pre-existing conditions can result in what is referred to as job lock workers remaining in jobs that may be a poor fit for fear of losing health insurance benefits should they move to another job. On the margin, this problem reduces the efficiency of labor markets, which depend in part on the absence of barriers to labor mobility. Therefore, the elimination of pre-existing conditions has the potential to boost economic growth in sectors outside of the health care market. Incentives for Primary Care. Considering these changes, it is clear that ACA will greatly increase the demand for primary care services. However, to counter the anticipated increase in demand, the ACA takes a few steps to increase the availability of primary care services for patients. Medicaid payments to family practitioners, general internists, and pediatricians will be 4 The Louisiana Medicaid physician fee schedule for CPT (doctor visit with existing patient) is $43.43 for patients over age 15 or $52.11 for ages 0-15 (Louisiana Department of Health and Hospitals, 2012). The Louisiana Medicare physician reimbursement for CPT is $49.16 for physicians practicing in New Orleans and $47.16 elsewhere in Louisiana (American Medical Association, 2012) 10 LSU E. J. Ourso College of Business Administration

15 2012 Health Occupations Outlook Table 2.1: Literature on Crowd-Out Article Data Source Methodology Crowd-out Definition Results Cutler and Gruber (1996) CPS Instrument eligibility with simulated eligibility based on entire nation; control for state, year, age; consider family level spillovers Dubay and Kenney (1996) 1988 and 1993 CPS Change in insurance coverage of children relative to change for adult men Dubay and Kenney (1997) 1988 and 1992 CPS Change in insurance coverage of pregnant women relative to change for men Thorpe and Florence (1998) NLSY Measure movement from private insurance onto Medicaid among children with privately insured parents Blumberg et al. (2000) 1990 SIPP Panel Compare change in insurance coverage of children made eligible by expansions to those not made eligible Yazici and Kaestner (2000) 1988 and 1992 NLSY Compare change in insurance coverage of children becoming eligible to those not becoming eligible Aizer and Grogger (2003) CPS Compare change in insurance, for those above AFDC eligibility vs below, in states with adult expansion, before vs. after expansion Card and Shore-Sheppard (2004) LoSasso and Buchmueller (2004) SIPP Panels Compare changes in insurance coverage of children around income and age limits for eligibility CPS Instrument eligibility with simulated eligibility based on entire nation; control for state, year, age, state x year; interact with state waiting periods Shore-Sheppard (2005) CPS Same as Cutler-Gruber, but add additional controls - children only Ham and Shore-Sheppard (2005) SIPP Instrument eligibility with simulated eligibility based on all other states: control for state, year, age Hudson et al. (2005) MEPS Compare changes in children made eligible and remaining ineligible; instrument with simulated eligibility Source: Gruber and Simon (2008) (Private insurance/public insurance) or (1-(uninsured/ public insurance)) (Private insurance/public insurance) (Private insurance/public insurance) % of those entering Medicaid with privately insured parents % of children made eligible losing private relative to gaining public (1-(uninsured/public insurance)) or (private insurance/public insurance) Coefficient on private coverage equation (no crowdout calculations) (Private insurance/public insurance) (Private insurance/public insurance) (1-(uninsured/public insurance)) or (private insurance/public insurance) (Private insurance/public insurance) (Private insurance/public insurance) Children 31%, or children: 40%, family level: 50% Below poverty: 15%, %: 22% Below poverty: 0%, %: 27% %: 59% 16% 4% 55-59%, 5-24% Statistically insignificant effect on private coverage for mothers and for children Below poverty, eligible for <100: 0; below poverty, eligible for : 50%; : 0 Average: 50% varies with state waiting periods 33% (age/year controls) to 59% (all controls), 0 No crowd-out Comparison: 25-55%, IV: 39-70% Division of Economic Development 11

16 Primary Care Physicians Considering some of the lessons learned through the previous research, Gruber and Simon also undertake a new study using an approach that addresses many of the pitfalls of previous efforts. The central finding is that crowd-out is more significant than these studies have suggested, especially when considering eligibility of all members of the family rather than only that of an individual. The authors estimate crowd-out to be on the order of 60%, saying that 60% as many individuals drop their private insurance as gain public insurance. Because health premium structures differ for individuals and families and purchases are typically coordinated within the family, family eligibility is extremely important to accurately measuring crowdout because most insurance carriers provide family plans with relatively low per-person costs if only a small portion of the family is eligible. Moral Hazard and Health Insurance. In the context of health care, moral hazard may lead those previously without health insurance to begin consuming higher levels of medical care after gaining coverage. Understanding the magnitude of this impact is important to anticipating how expanding coverage in the coming years will impact the demand for health care professionals. Empirical research indicates that moral hazard has a large impact on demand for care with uninsured individuals receiving approximately half of the care of insured individuals on a person basis (Hadley & Holahan, 2004). Therefore, newly insured individuals would be expected to almost double consumption of medical services after becoming insured leading to a larger increase in the need for workers than would otherwise be expected. Impact of Adverse Selection on Demand for Care. Another issue that must be considered to correctly gauge the impact of the expansion of coverage due to ACA is the level of care that will be demanded by the newly insured. Economists have studied the issue of adverse selection in insurance purchasing to assess whether or not individuals who expect to demand fewer services are less likely to buy care. If this effect is large, then the impact of expanding coverage will be muted because the newly insured will have much lower demand for services relative to those who are currently insured. However, the literature has generally found little evidence that adverse selection plays a major role among individuals. 5 A recent study by Krueger and Kuziemko (2011) looked specifically at the question of how significant adverse selection might be among those impacted by the ACA. The study found no evidence of adverse selection or a difference in costs of medical services between previously uninsured and insured individuals. 3. Primary Care Physicians Background Structural issues have depressed the relative number of primary care physicians, a problem in the provision of quality health care that may be exacerbated by the ACA. This section discusses the reasons why the supply of primary care physicians is relatively lower today than in the past, and offers potential solutions. There are many reasons for this decline, including a lack of interest in primary care among medical students and the overall difficulty in becoming a physician. Furthermore, certain populations, such as people covered by Medicare and Medicaid, and those living in rural areas, find it more difficult to access a primary care physician. Limited access to primary care services also extends to other sectors of the patient population. Schoen et al. (2010) found that patients often have extended waits for accessing primary care services, with one in five adults reporting a delay of at least 6 days. Pitts et al. (2010) found that 46 percent of ER patients in their sample would have preferred to see a primary care physician, but were unable to obtain an appointment. The lack of interest in primary care among medical students is at least in part due to primary care physicians lower earnings relative to specialists. Adjusted for inflation, salaries of primary care physician have been declining while the demand for their services has been rising (The Commonwealth Fund, 2001). Colwill et al. (2008) predict that overall growth in the population along with an aging population will increase the demand for primary care services resulting in a concomitant increase in the primary care physician s work load by a third between 2005 and While increasing demand and falling salaries seem contradictory at first blush, this odd result is largely due to the way that public and private health insurance payments are structured. Insurance companies and public programs such as Medicare generally pay physicians for providing treatments, and do not always pay for primary care services such as taking patient histories, performing examinations, or coordinating care across multiple organizations. The ACA took a few steps to address this problem by increasing payments to primary care physicians (The Commonwealth Fund, 2001). We address this issue in more detail below. Becoming a physician is difficult and costly due to high barriers to entry caused in part by strict admissions requirements for medical school, lengthy training requirements, high cost of medical school (leaving medical school graduates deeply in debt), and the limited number of slots for residencies required in order to practice medicine. The relatively large debt physicians incur during medical school influences physicians choice of specialties. Other things equal, medical students have an incentive to opt for the higher-paid specialties instead of primary 5 A couple of recent examples include Diehr (1993) who looked at a subsidized health insurance plan in Washington State and Decker et al. (2011) who compare new Medicare enrollees who had insurance to those who did not have insurance just before enrolling in Medicare. 12 LSU E. J. Ourso College of Business Administration

17 2012 Health Occupations Outlook Figure 3.1: Health Professional Shortage Areas of Louisiana, 2011 Source: US Census Bureau Department of Health and Human Services care. The limitations on medical school enrollments and the limited number of residency positions available help to ensure quality physicians, but relaxing these constraints could be one option for increasing the supply of primary care physicians. Again, the ACA took steps to increase the future supply of primary care physicians by increasing residency positions for and redistributing unused residency position toward primary care providers (The Commonwealth Fund, 2001). Patients living in rural areas have more trouble accessing primary care than those living in urban centers. Since many physicians prefer to live in urban or suburban areas, rural patients areas are more likely to have limited access to primary care. Reschovsky and Staiti (2005) reported that difficulties in accessing medical care in rural areas are due in part to the disparity in the number of physicians per capita between rural and urban areas. For example, the authors report that in rural areas there are on average of 53 primary care physicians per 100,000 people, compared with 78 primary care physicians per 100,000 people in urban areas. This disparity is even greater for specialists 54 versus 134 per 100,000 people, respectively. A health professional shortage area (HPSA) for primary care is defined as a region that has more than 3500 people per primary care physician or more than 3000 people per primary care physician but unusually high needs for primary care (Health Resources and Services Administration, 2012). About 65 million Americans live in an area with a shortage of primary care physicians (Kaiser Commission on Medicaid and the Uninsured, 2011). Although rural physicians tend to have slightly lower incomes, it is unlikely to be the chief reason for the observed shortage. Because price levels are lower in rural areas, rural physicians actually have higher incomes after adjusting for cost of living (about 13 percent higher). Figure 3.1 shows the primary care HPSAs in Louisiana. Strategies for increasing rural access to primary care include increasing the total number of primary care physicians, creating programs that make rural areas more attractive to primary care physicians (e.g. loan forgiveness programs), and expanding the role of nurse practitioners and physician assistants. These medical professionals require less training and do not command the same level of compensation as physicians. But, expanding their role in the provision of health care is controversial because of concerns about lowering the quality of health care. Currently, Louisiana is one of 27 states that require a physician s involvement for both diagnosis and treatment (Kaiser Commission on Medicaid and the Uninsured, 2011). Division of Economic Development 13

18 Primary Care Physicians While geography can make it more difficult for some patients to access primary care, an insufficient number of primary care physicians might also disproportionately impact patients with some types of health coverage. Medicaid and Medicare patients could also have difficulty seeing a primary care physician because both Medicare and Medicaid pay lower reimbursement rates than most private insurance plans. Furthermore, because Medicaid reimbursements are lower than Medicare, that group will likely see the most restricted access to care. In 2008, Zuckerman, Williams, and Stockley (2009) found that physician fees for primary care visits in Medicaid fee-for-service programs were on average only 66 percent of the equivalent fees in Medicare and the Medicaid/Medicare fee disparity varied widely from state to state. The authors reported that the Medicaid-to- Medicare fee ratio ranged from 36 percent in Rhode Island to 1.40 in Alaska. Louisiana s ratio was 90 percent. It is important to note that any future shortage in primary care physicians would likely hit Medicaid and Medicare patients first. Measuring Employment Because physicians typically work long hours, but vary the number of hours in predictable ways throughout a career, having a good understanding of the availability of primary care physicians requires considering both the number of physicians and the hours worked. However, developing accurate measures of the number of physicians or their hours worked is not a straightforward task. Physicians are not captured well by many common measures of employment for several reasons. A central reason for this is the prevalence of self-employed physicians who are not captured by employer-based surveys. According to the American Community Survey, 36.4 percent of physicians in Louisiana are self-employed. Moreover, because physicians work long hours and earn such high wages, it is more difficult to encourage their participation in surveys. To overcome these obstacles and develop an accurate measure of physician labor supply, we consider licensing data in addition to standard employment statistics as well as the results of an original survey conducted as part of this study. Occupational Employment Statistics One of the most common and widely cited measures of employment is derived from the Occupational Employment Statistics (OES) program. The Bureau of Labor Statistics utilizes a semi-annual survey to collect data on all nonfarm wage and salaried workers in the U.S. The OES sample excludes individuals who are self-employed, and owners and partners in unincorporated businesses. Because a significant number of physicians likely fall into this category, these estimates may underrepresent the true number of physicians working at a point in time. Though incomplete, the data provide useful comparisons over time and geography. For this study, the specific occupations of interest are those most highly associated with primary care: family and general practitioners, general internists, pediatricians, and obstetricians and gynecologists. According to estimates from the OES survey, Figure 3.2 shows how the supply of primary care physicians has varied over the last decade in Louisiana and the nation. In 2001 there were approximately 6.8 primary care physicians per 10,000 population in Louisiana compared to a little over 9.8 per capita nationally. Louisiana had roughly 31 percent fewer physicians per capita than the national average. Between 2001 and 2011 the supply of primary care physicians per capita fell in both the state and the nation. By 2011, per capita supply of primary care physicians in Louisiana had fallen to 5.2 (roughly a 24 percent reduction), in the per capita supply in the nation fell to 7.7 (a 21 percent reduction). Louisiana now has roughly a third fewer primary care physicians per capita than the national average according to OES. Licensing Data Since state law requires a license for physicians to practice medicine, licensing data provide a fairly accurate count of the supply of physicians across specialties within the state. However, licensing data include both physicians who are active and those that are inactive but choose to maintain their licenses. It also includes physicians who move out of Louisiana or work in another profession yet elect to maintain a Louisiana license. Physicians who are not engaged in direct patient care, such as administrators, researchers or educators may report on their license form that they are actively engaged in the profession. From the perspective of measuring the supply of physicians actively engaged in patient care within the State, the licensing data may slightly overstate the effective supply. With these issues in mind, an original survey was developed to help identify active physicians among those with Louisiana licenses to provide the most accurate assessment of physician labor supply. Number of Primary Care Physicians There are significant differences between estimates of the number of primary care physicians produced by licensing data and OES data. For this comparison we use published Kaiser Family Foundation estimates from state licensing data and our per capita estimates using the OES data. It is important to point out that the Kaiser Family Foundation includes the following fields in its definition of primary care: internal medicine; family medicine/general practice; obstetrics/gynecology; and pediatrics. This list of specialties provides the most direct comparison between the two data sources. These comparisons are based on 2008 data, the only year we have comparable data from both sources. Per capita supply estimates are shown in Table 3.1 for Louisiana and 11 other southeastern states as well as the nation. The data show that the number of primary care physicians based on licensing data exceeds supply estimates from the OES survey data by a wide margin. Nationally, the OES survey estimates 14 LSU E. J. Ourso College of Business Administration

19 2012 Health Occupations Outlook Figure 3.2: Total Primary Care Physicians per 10,000 Individuals, Louisiana and United States Source: Calculated from Occupational Employment Statistics (OES), (Bureau of Labor Statistics, 2012b) fewer primary care physicians per 10,000 population than the estimates coming from licensing data. Second, it is interesting to note that the difference between the per capita supply estimates from the two sources of data is larger in Louisiana (19.7) than it is for most southern states, exceeded only by Florida (22.0). Table 3.2 shows the difference in absolute numbers between the OES estimates for primary care physicians from 2011, the most recent year available, and a summary of licensing data from 2012 compiled by the Kaiser Family Foundation (2012a). Note that the OES category and license counts reported by the Kaiser Family Foundation for general surgeons is simply titled surgeons which does not exactly match the category used for this study, general surgery. While these two sources produce significantly different results, we have more confidence in developing a count of primary care physicians using the licensing data. The aforementioned limitations of the data can be overcome using some screening of the license data to account for out-of-state physicians as well as the survey results, which help fill gaps in reported specialty from the licensing data and account for retired or nonpracticing physicians. The process of developing this estimate is discussed in greater detail in the context of the survey results presented below. Physician Hours Not only is it important to have a clear sense of how many physicians are practicing, but it is also helpful to have a clear sense of how many hours these physicians can be expected to work today and in the future. A study by Staiger et al. (1999) found that the average number of hours worked per week by physicians in the U.S. fell by about 7 percent between 1996 and A very similar pattern existed in Canada. Crossley et al. (2006) found that between 1981 and 2001, physicians in that country worked 5 fewer hours per week, on average. The widespread reduction in the time physicians spend practicing medicine is contributing to concerns about the possibility of a shortage of physicians, particularly with respect to primary care physicians. One possible explanation for this reduction in work effort is that physician fees have been falling. For example, Staiger and his colleagues found that in the U.S. physician fees decreased by 25 percent in real terms between 1995 and Division of Economic Development 15

20 Primary Care Physicians Table 3.1: Primary Care Physicians per 10,000 Civilian Population, 2008 Licensing Data OES Difference National Southeast Virginia Tennessee Louisiana Florida North Carolina West Virginia Kentucky South Carolina Alabama Georgia Arkansas Mississippi Source: Kaiser Family Foundation (2012a), Bureau of Labor Statistics (2012b), authors calculations Table 3.2: Number of Primary Care Physicians in Louisiana Practice Area Licensing Data 2012 OES 2011 Family and General Practice 1,403 1,200 Internal Medicine 1, Obstetrics and Gynecology Pediatrics General Surgery Total 5,662 2,380 Source: Kaiser Family Foundation (2012a), Bureau of Labor Statistics (2012b) 16 LSU E. J. Ourso College of Business Administration

21 2012 Health Occupations Outlook Key Changes in the Health Care Industry Medicare. Medicare is the federal health security program created in 1965 as part of the Social Security Act. Only American citizens 65 and older or those with certain disabilities are eligible for Medicare. Medicare has four components, with Medicare Part B, which covers outpatient care, being most relevant to primary care. Medicare covers most of the cost of health care for recipients participants are required to pay a portion, typically about 20 percent (Barry, 2012). This section will focus on the details of Medicare that are relevant to this study; namely how the Medicare program reimburses health providers, and, more importantly, at what rate. Medicare reimburses health providers at a set rate depending on the medical condition and treatment plan. In 1997, Congress included a provision in the Balanced Budget Act that fixed Medicare physician payment rates to a formula based on economic growth called the sustainable growth rate (Kaiser Health News, 2012). While the formula gave physicians higher payment rates for the first few years, in 2002 the sustainable growth rate began calling for payment reductions. However, in that year Congress postponed the cuts and has continued to do so ever since. The gap between current reimbursement rates and the rates mandated by the 1997 legislation has widened over time. If Congress failed to postpone the automatic rate changes (which the media has dubbed the doc fix ), physicians would face a 27 percent cut in Medicare payments (Kaiser Health News, 2012). While Congress has so far acted to postpone these cuts, removing the sustainable growth rate provision entirely would clearly result in a Medicare program far more expensive than projected, so the result has been a series of short-term postponements. The most recent postponement of the reimbursement cuts mandated by the sustainable growth rate was this past February. To help estimate how significant the impact of such a change would be to access, we incorporated a series of questions about acceptance of payment types among physicians to physician survey. Medicaid. Created simultaneously with Medicare, Medicaid is a federal health insurance program for low income individuals without access to private insurance. It is also partially funded and almost entirely managed by state governments. Each state determines its own eligibility requirements for Medicaid recipients as well as how much its program will pay for various services. Hence, Medicaid programs vary between states. Medicaid in Louisiana is undergoing a transformation to the Bayou Health Network, which is a partial privatization of the Medicaid program. Five different companies will act as Coordinated Care Networks (CCNs), which will compete for Medicaid participants and actively manage their health. The reform is estimated to save about $135 million, but also improve care by shifting 800,000 people and $2.2 billion in public spending to private companies (Barrow, 2012). The five CCNs are Louisiana Healthcare Connections, LaCare, Amerigroup, UnitedHealth, and Community Health Solutions. While the latter two CCNs will have their expenses reimbursed by the state using the traditional fee-for-service model, the first three CCNs will instead get a fixed monthly payment from the state government for each Medicaid recipient that signs with that company. In the coming years, the Affordable Care Act has the potential to bring about a second major change to Medicaid by significantly expanding coverage. Cunningham (2011) estimates that Medicaid enrollment will expand by as many as 16 million people nationally. Holahan and Headen (2010) in turn, estimate that new Medicaid enrollees in Louisiana would be somewhere between 350,000 and 507,000, depending upon the participation rates of the newly eligible uninsured. However, Louisiana is one of several states that is currently planning not to expand coverage in response to the ACA. Should such an expansion eventually occur, there likely would be concerns about the current capacity of health care delivery systems, in particular the supply of primary care physicians. One unintended consequence of a reduction in the financial barriers to medical care for the newly insured would be the emergence of other barriers to care, such as the availability of primary care physicians and a patient s ability to access care in a reasonable amount of time. Unless there is a sufficient increase in the supply of primary care physicians willing to treat Medicaid patients, long queues would likely develop. The chief reason that supply constraints arise from the expansion of the Medicaid eligible population is that more physicians accept privately insured or Medicare patients than accept Medicaid patients, primarily due to the lower reimbursement rates under Medicaid (Cunningham P., 2011). To address this problem, the ACA incorporates a temporary increase in Medicaid reimbursement rates for primary care. However, this rate increase is limited to 2013 and Clearly, the fact that this rate increase is temporary will dampen physician incentives to accept more Medicaid patients. However, there are other provisions in ACA that are designed to increase the availability of primary care. Accountable Care Organizations. The emergence of Accountable Care Organizations (ACOs) is a major development within the health care industry and may become an increasingly important segment of this industry. While the definition of an ACO is somewhat murky, one workable definition is: a network of doctors and hospitals that shares responsibility for providing care to patients (Kaiser Health News, 2011). The goal of an accountable care organization is to coordinate the care of patients across multiple medical organizations and types of physician practice. This should help to reduce costs by eliminating duplicated or unneeded procedures. As a large medical organization that attempts to reduce costs, ACOs seem Division of Economic Development 17

22 Primary Care Physicians similar to HMOs, but ACOs generally have a greater emphasis on giving patients more options than they would have in an HMO. Insurance companies, hospitals, and groups of physicians have all attempted to form ACOs. As the prevalence of these organizations increases, so too will the demand for primary care physicians to help coordinate care. Patient-Centered Medical Home. The patient-centered medical home (PCMH) movement is another trend in health care that focuses on primary care. The ACA roughly defines a medical home as a mode of care that includes personal physicians, whole person orientation, coordinated and integrated care, safe and high-quality care through evidenceinformed medicine, appropriate use of health information technology, and continuous quality improvements, expanded access to care, and payment that recognizes value from additional components of patient-centered care (Government Printing Office, 2010). This complex definition underscores the idea that a patient-centered medical home is not really one concept but a collection of many concepts. Furthermore, since medical organizations can adopt some but not all of these qualities, there are actually varying degrees of medical homeness. (Robert Wood Johnson Foundation, 2011) In essence, a patient-centered medical home is similar to an accountable care organization except that a medical home is a single primary care provider rather than an organization with multiple types of health providers. The motivations for primary care providers to become a medical home are twofold: increase patient care and reduce expenses. Probably the most important reform of medical homes is that the primary care physician is responsible for coordinating all of a patient s medical care. Medical home physicians also attempt to be more accessible by communicating with patients electronically and having more evening and weekend hours. Medical homes save money in much the same way as ACOs, which is by managing a patient s care so as to reduce duplication or unnecessary procedures. Also, the intention is for the primary care physician to develop a closer relationship with patients, which may allow for a greater emphasis on preventative care. The American Academy of Pediatrics first proposed the concept of a patient-centered medical home in 1967 as a program for children with special health care needs (Robert Wood Johnson Foundation, 2011), but the idea has gained support for broader implementation. Although medical homes are not a major focus of the ACA, it did allocate funds for medical home pilot projects. We expect the medical home concept to continue to grow, affecting primary care physicians in multiple ways. Since primary care physicians are the leaders of all medical homes, the PCMH movement could increase the demand for primary care physicians, but the greater emphasis on primary care that medical homes create might also increase the attractiveness of primary care to medical students. Review of Relevant Literature Elasticity of Physicians Labor Supply. We now consider how responsive physicians supply decisions are to changes in their remuneration. If monetary considerations are important for labor supply decisions by physicians, then labor shortages in this market for a particular specialty or group of specialties may be relatively short-lived. That is, if money matters, and reimbursement rates increase for a particular specialty, physicians in that specialty will respond by working more, some physicians in other specialties will shift into that specialty, and more medical students will select training in that specialty. Thus, there is both a short run response (physicians already in practice who are willing to work more) and a longer run response (physicians who change specialties and the increase in the number of medical school graduates in that specialty). Understanding physician responses to changes in their income levels (e.g. increased Medicaid reimbursement rates) is a crucial component of this project. A number of provisions in the ACA change reimbursement rates, such as raising Medicaid rates to Medicare levels in 2013 and 2014 for primary care services. In order to forecast the supply of primary care physician services over this decade, it is important to know how primary care physicians will respond to changes in their remuneration that is, how elastic is the supply of labor in this market? Economic theory strongly suggests that labor supply is more elastic (more responsive) for lower-skilled and lower paid professions than it is in high-skilled occupations. For example, among child-care workers, estimates of the price elasticity of supply were found to be about 2.0 (very price elastic). What this suggests is that a 10 percent increase in their hourly wage results in a 20 percent increase in the quantity of labor supplied (Blau, 1993). But, among workers in a more highly-skilled profession, like nursing, the estimated labor supply elasticity was found to be approximately 1.1 (Burkett, 2005). So, a 10 percent increase in their pay would lead only a 1 percent increase in the quantity of labor supplied. As expected, supply elasticity estimates within lower-paying occupations contrast sharply with the supply elasticity estimates for physicians. In this section we review the major empirical literature regarding the size of the labor supply responses to a change in the physician wage and whether the supply curve for physicians is backward-bending. In most of the existing literature dealing with the estimation of labor supply elasticities for physicians, there are two primary targets, studies that examined physician service supply and studies concerned with estimating physician labor supply (e.g. quantity of physicians in the market). In other words, in response to an increase in demand for primary care services (and resulting increase in fees) there are at least two margins for adjustments to take place changes in the services supplied by existing physicians and secondly, changes in the number of physicians available in the market. 18 LSU E. J. Ourso College of Business Administration

23 2012 Health Occupations Outlook Table 3.3: Physician Labor Supply Elasticities Study Physician Category Elasticity Estimate Rizzo & Blumenthal (1994) Young Physicians (<40) 0.3 Female Physicians 0.49 Male Physicians 0.23 Brown (1994) Primary Care Physicians 0 Bradford & Martin (1995) General Practice 0.02 Family Practice 0.02 Pediatrics 0.02 All Other Specialties < 0.0 Showalter & Thurston (1997) Self-Employed Physicians 0.3 Thornton (1998) Physician Services 0.06 Saether (2005) Norwegian Physicians 0.18 Baltagi et al. (2005) Norwegian Physicians The seminal work on the theoretical foundation for supply elasticity among physicians is McGuire and Pauly (1991). The authors examine the theoretical case of physicians facing multiple payers, which characterizes much of the American health system (with private insurance, Medicare, Medicaid, and other miscellaneous payers). McGuire and Pauly divide physician responses into two groups: those who maximize profits and those who target income. Profit-maximizing physicians respond to a reduction in payments for a service by providing less of that service. This follows from standard economic theory of profit-maximizing firms. If the selling price decreases, assuming increasing marginal costs, the firm will respond by reducing production. However, physicians can react in the opposite manner if they practice income targeting. In this case, the physician s marginal utility of income is so high that they respond to a reduction in payment by providing more services, so as to maintain their target income level. In the profit-maximizing case, physicians will substitute other types of patients (e.g. privately insured patients for Medicare patients) or substitute leisure for labor. In the incoming targeting case, physicians will potentially increase their services for multiple payment types, while decreasing their leisure time. Most of the early studies found evidence that physicians labor supply curve was backward-bending physicians provide more services when payments decrease. 6 If they are indeed backward-bending, they could reduce the expected benefits of policies intending to control medical expenditures if physicians respond to reduced reimbursement per visit by encouraging more visits. Dafny (2003) also found evidence of similar behavior on the part of hospitals. Many of these studies found that physicians are more likely to increase the use of tests and procedures that are more profitable to them. This is consistent with the theory of induced demand (physicians create demand for their services). More recent studies have found evidence that physicians generally do not engage in income targeting behavior, but instead attempt to maximize profits. 7 These studies generally 6 See Christensen (1992), Mitchell and Cromwell(1995), Nguyen and Derrick (1997), and Yip (1998). 7 See Hadley, Mitchell, and Mandelblatt (2001), Mitchel, Hadley, and Gaskin (2002), Hadley et. al. (2003), Hadley and Reschovsky (2006), and Hadley, Reschovsky, Corey, and Zuckerman (2010). Division of Economic Development 19

24 Primary Care Physicians argue that income targeting and demand inducement do occur, but that the practice is rare and largely confined to physicians who have a relatively large Medicare patient practice. Thus, the important question is whether income targeting is widespread enough for the elasticity of supply for physicians in a large sample to be negative, and these authors generally do not think so. A survey of physicians supply behavior by Nicholson and Propper (2012) provides an excellent summary of the emerging consensus among the most recent studies in this area. Their synthesis of the literature concludes that physicians are not particularly responsive to wage changes and the effects on their labor supply decisions are small that is, the income elasticities are small. In the most recent studies reviewed by Nicholson and Propper the wage elasticity estimates ranged from zero to 0.3 with the bulk of the estimates ranging from 0.18 to Generally, studies find higher supply elasticities for primary care physicians than for specialists. Also, self-employed physicians, female physicians and younger physicians exhibit higher supply elasticities than their counterparts. Table 3.2 provides a summary of the elasticity estimates produced in the more recent empirical literature (post 1980s). Whereas the early literature used aggregate time series data on physician services and fees to estimate supply elasticities, more recent studies employed micro-level data on individual physicians. Two of these studies focus on primary care physicians specifically. While Brown (1994) found the supply elasticity of primary care physicians to be insignificantly different from zero, Bradford and Martin found significant positive supply elasticities for these specialties. In fact, their supply elasticity estimates for three groups of primary care services were the same (0.02). Considering the variations across studies in methodologies, data sources, groups of physicians sampled and countries studied, it is safe to conclude that there is general agreement within the more recent literature that physicians are not highly responsive to wage changes. But, across all specialties, primary care physicians tend to be more responsive (higher elasticities) than other physicians. Therefore, in forecasting future supply of primary care physicians and primary care services, it would be unwise to ignore the positive supply elasticity that exists for this group of physicians. Our review of the literature suggests that a reasonable range for the supply elasticity of primary care physicians is somewhere in the range of 0.18 to For the purposes of this study it is critical to understand how to interpret these supply elasticities. Technically, these measures of responsiveness to changes in remuneration (wages, salaries or reimbursement levels) assume a permanent one-time change in payment levels. The underlying behavioral response of the physician is conditioned on the physician assumption that the change is not temporary. Economic theory strongly suggests that a temporary change in remuneration elicits a smaller change in supply than a permanent change in remuneration. In effect, the range of elasticity estimates summarized above is an upper bound for the supply elasticities that would apply to the reimbursement changes incorporated in the ACA. Recall, the ACA raises Medicaid reimbursements to Medicare levels for only two years 2013 and It remains unclear whether these reimbursements will become permanent, disappear after the first two years, or fall somewhere in between. Of course, supply decisions by physicians will be determined on their perceptions of these future payments. Thus, we expect the supply elasticity of primary care physicians to be on the lower end of our preferred range. Physician Participation in Medicaid Programs. While current Louisiana policy makers would not implement the Medicaid expansion, the impact of such a change was considered to help inform future decisions. A major concern about the potential expansion of Medicaid eligibility is that a significant portion of physicians do not currently participate in Medicaid. Therefore, it is possible that the increase in Medicaid eligibility will not secure health care for low-income individuals. As mentioned in the section concerning the Affordable Care Act, this health reform legislation not only would expand Medicaid but also increase Medicaid reimbursements for primary care providers to Medicare levels in 2013 and The most comprehensive study dealing with physician acceptance of Medicaid patients is Sloan, Mitchell, and Cromwell (1978). They used survey data from fee-for-service physicians to estimate the relationship between a physician s willingness to accept Medicaid and other explanatory variables. Not surprisingly, higher Medicaid reimbursements were associated with greater willingness to accept Medicaid patients. However, the authors also found that administrative burdens of Medicaid negatively impacted Medicaid acceptance. These administrative burdens included time-consuming forms, forms that were returned unaccepted, and significant delays before payments were received. These issues are apparently more severe for Medicaid than Medicare, and present an additional challenge for the potential expansion of Medicaid. Other papers that probed the issue of physicians acceptance of Medicaid patients found similar results. 8 These papers also found a positive relationship between higher reimbursements and acceptance rates for Medicaid patients. Perloff et al. (1995) argued that the effect of Medicaid reimbursement rates on Medicaid acceptance was so small that raising Medicaid reimbursement rates to Medicare levels would not provide Medicaid patients with the same access to physicians as Medicare patients. Although there may be other reasons for a physician s unwillingness to accept Medicaid patients, Sloan, 8 See Perloff, Kletke, and Neckerman (1987), Mitchell (1991), and Perloff, Kletke, and Fossett (1995). 20 LSU E. J. Ourso College of Business Administration

25 2012 Health Occupations Outlook Mitchell, and Cromwell argue that administrative hassle is an important component. Cunningham and O Malley (2008) look solely at the relationship between reimbursement delays and Medicaid participation. They note that Medicaid has much longer reimbursement turnaround than other forms of payment, thus reducing a physician s willingness to accept Medicaid patients, particularly for more expensive treatments. The time value of money is important here, as many physician practices are essentially losing money while payment requests are tied up in the Medicaid reimbursement system. If these administrative burdens are not addressed, then expanding Medicaid eligibility and increasing Medicaid reimbursements may not be enough to ensure that low-income individuals have access to primary care. 4. Clinical Laboratory Personnel Background Clinical laboratory personnel are essential to much of the testing that occurs behind the scenes to provide critical information to physicians in their diagnoses and treatment decisions. An efficient clinical laboratory relies on the right combination of workers with differ levels of education and training to handle tasks of varying degrees of complexity. As in other areas of the labor market, there is a general expectation that workers with more education or training can provide higher quality services. However, in return for that perceived improvement in quality, workers with more education or training typically demand higher wages. Clinical laboratory personnel have been included in this year s study because of recent efforts to require higher levels of education for these workers in Louisiana. To determine the benefits of such a regulation, the potential costs must be weighed against the potential benefits. This study focuses primarily on determining the demand for a more educated workforce by Louisiana employers. Measuring Employment We obtained data on the total number of clinical laboratory personnel in Louisiana from both the Louisiana State Board of Medical Examiners (LSBME) and the BLS Occupational Employment Statistics. Table 4.1 shows the number of clinical laboratory technologists and technicians in Louisiana based on those two sources. The OES estimates come from May 2011 and are based on surveys conducted from 2008 to 2010, while the licensing counts are drawn from administrative records in January While the OES categorizes occupations by the Standard Occupational Classification (SOC) System, the licensing board groups clinical laboratory personnel into generalists, specialists, cytotechnologists, technicians, laboratory assistants, and phlebotomists. Based on conversations with the LSBME as well as referencing the BLS Occupational Outlook Handbook, we determined that the generalists, specialists, and cytotechnologists are best classified in the SOC code for clinical laboratory technologists ( ). Technicians and laboratory assistants in the licensing data are best classified in the SOC code for clinical laboratory technicians ( ). Phlebotomists have their own SOC code, Thus, we included the totals of the technologist and technician occupations in the table and omitted phlebotomists altogether. Overall, the two sources show similar numbers of technologists while the OES estimates have significantly more technicians than are present in the licensing data. State law offers an exemption to licensure for technicians in certain clinical labs including those operated by nonprofit entities for instruction and research as well as those used for forensic testing in the criminal justice system (Louisiana Revised Statues, 1993). This exemption is likely to explain much of the difference between the two employment measures. Because the OES measure provides a broader measure of the workforce needs for these types of workers, this study relies on those estimates as a baseline measure for employment. Table 4.1: Comparison of Clinical Laboratory Personnel Counts Occupational Employment Statistics (2011) Clinical Laboratory Technologists ( ) 2,870 3,001 Clinical Laboratory Technicians ( ) 2,320 1,331 Total 5,190 4,332 Louisiana State Board of Medical Examiners (2012) Source: Bureau of Labor Statistics (2012b), Louisiana State Board of Medical Examiners Division of Economic Development 21

26 Clinical Laboratory Personnel/Medical Coders Effect of Minimum Education Requirements Currently, most clinical laboratory technologists have bachelor s degrees with courses in chemistry, biology, microbiology, mathematics, and statistics, while most clinical laboratory technicians have associate s degrees with courses in science and clinical laboratory science. (Bureau of Labor Statistics, 2012c) However, only 71 percent of technologists have at least a bachelor s degree (O*NET Online, 2012a) and only 86 percent of technicians have at least an associate s degree (O*NET Online, 2012b). Currently in Louisiana, these individuals must be licensed by the Louisiana State Board of Medical Examiners, but the licensing board does not require bachelor s degrees for technologists and associate s degrees for technicians. Instead of completing the respective degrees, these individuals only need to apply for and successfully complete any nationally recognized technologist or technician certification. A bill was proposed in the 2012 legislative session that would have required laboratory workers to hold a bachelor s degree to obtain a clinical laboratory technologist license and an associate s degree to obtain a clinical laboratory technician license in the state (Louisiana State Legislature, 2012). While this bill failed to garner enough votes this pass in the recent legislative session, a similar bill could likely be reconsidered in the future. Therefore, this study assesses the viewpoint of employers toward the adequacy of education among their workers. While not a complete study of the benefits or costs of such a proposal, it is a relevant indicator of the value of such a change as employers will likely appreciate the potential for increasing quality as well as the potential for increasing costs. Finally, it is worth noting that technologists without at least a bachelor s degree and technicians without at least an associate s degree are likely to be older employees, and any education requirements would probably not affect them but instead only apply to new applicants for state licensure. In theory, education requirements would make it more difficult to join the ranks of an occupation, would reduce the labor supply of that particular skill, and thus drive up the wages of the members of that occupation. In a study of radiologic technologists, a similar allied health profession, Timmons and Thornton (2007) found that licensing (which generally included an education requirement) of radiologic technologists was associated with a 3.3 percent increase in radiologic technologist wages. White (1978) studied clinical laboratory personnel and found that licensing requirements in San Francisco and Los Angeles increased clinical laboratory technologist wages by percent. However, the education requirements in this case were severe, mandating that all clinical laboratory personnel have bachelor s degrees, which essentially mandates that all technicians be replaced with technologists. Still, there is considerable evidence that education requirements will increase wages and reduce the supply of the affected workers, which would lead to higher health care costs and possibly even labor shortages. 5. Medical Coders Background Medical coders play an important role in keeping the health care system running with responsibility for categorizing diagnoses and treatments for records and billing purposes. They are a third group of workers facing significant changes in the coming years with the move to a much more complicated coding system on the horizon. These changes are significant enough to raise concerns about early retirements and younger workers leaving the field rather than pursuing the additional training required. To better understand the potential for workforce shortages, this study assesses employer perceptions regarding employee willingness to pursue training and continue working as medical coders. Measuring Employment Unlike physicians and clinical laboratory personnel, medical coders do not require licenses to work in Louisiana. Because of this, there are no licensing databases that have information about the number of medical coders in Louisiana. However, the Occupational Employment Statistics does contain an estimate of the number of medical coders. The most recent estimates from 2011 indicate that 2,360 medical coders work in Louisiana. International Classification of Diseases 10th Revision The upcoming adoption of the International Classification of Diseases 10th Revision (ICD-10) will largely affect medical clerks and other medical records employees. The current deadline for the adoption of ICD-10 is October 1, 2013, though the Department of Health and Human Services has postponed previous deadlines and there is a proposed rule under advisement that would delay the implementation one year further to October 1, 2014 (Centers for Medicare and Medicaid Services, 2012b). This classification system provides codes for various diseases and treatments. The medical community currently uses ICD-9, which has been the standard for about 30 years. The new ICD-10 will provide updates based on recent medical developments, expand the number of codes to provide much more detail, and split the codes into two sets: ICD-10- CM for diagnoses and ICD-10-PCS for procedures (American Hospital Association, 2009). The additional level of detail that should make ICD-10 much more useful in reporting diseases also makes ICD-10 much more difficult to work with. As such, many medical coders who work with ICD-9 need additional training to work with ICD-10. Medical coders generally have an associate s degree, but there is concern that many medical coders may not have had the necessary coursework in anatomy and physiology, though they are now generally considered part of a medical coder s training (Bureau of Labor Statistics, 2011b). Furthermore, many coders 22 LSU E. J. Ourso College of Business Administration

27 2012 Health Occupations Outlook may decide to change professions instead of incurring the full costs of retraining. Coders close to retirement may elect to retire early. This would have a harmful impact on employers of a profession that is already in high demand. In order to understand the magnitude of this potential problem, the study analyzes medical coders education, readiness for ICD-10, and retirement plans. Electronic Billing Systems and Electronic Medical Records Another major change within the health care industry is the transition to electronic medical records and electronic billing systems. On January 1, 2012 the Centers for Medicare and Medicaid Services began requiring all transactions to be electronic (Centers for Medicare and Medicaid Services, 2010). Although this change is required for only Medicare and Medicaid, the size of the Medicare patient population essentially requires almost all health organizations to convert their billing systems to an electronic format. The CMS reports that in 2010, 99 percent of Medicare Part A claims and 96 percent of Medicare Part B claims were processed electronically (Centers for Medicare and Medicaid Services, 2010), so it appears that most employers have already made the transition. There is also an upcoming deadline for health employers to adopt Electronic Medical Records (EMR), which, as the name suggests, means transferring all patient information and medical histories into an electronic format. The CMS deadline for this transition is 2015, with health employers facing penalties for not doing so in 2016 (American Psychiatric Association, 2012). Electronic medical records should permit patients to access their medical history online and make it easier for multiple physicians to have access to a patient s complete medical history. However, implementing electronic medical records is time-consuming, requires information technology expertise, and direct reimbursement for the provider s costs associated with making this conversion. 6. Health Works Surveys Overview The 2012 Health Occupations Outlook included two separate surveys, which were intended to capture different perspectives on future changes in the health care sector. A survey of employers captured the demand for workers as well as some key characteristics of health employers. A survey of physicians targeted questions about physician supply directly to Louisiana physicians in five occupation groups that focus on primary care. Both survey instruments are available in Appendix B. The administration of the two surveys was staggered because many small employers of primary care physicians, such as a single physician practice office, would likely get a copy of both surveys and questions from each were important for anticipating the demand and supply of workers at these types of facilities. To improve our response rates, we made both surveys available in mail, web, fax and phone versions. Employer Survey The employer survey went to health employers drawn primarily from Info-USA s employer database. In order to ensure that the state s largest health employers were represented in the survey, we supplemented the addresses and phone numbers from Info-USA with contact information from the Louisiana Hospital Association (LHA). To encourage participation by these large employers, we ed each hospital CEO with a copy of the employer survey, and then later called them directly to ask for their participation. In order to target the survey to establishments most likely to employ the kinds of workers included in this year s study, the sample consisted of all employers in 8 different North American Industrial Classification System (NAICS) categories that were expected to have the highest concentrations of primary care physicians or clinical laboratory personnel under the assumption Table 6.1: Establishments by Industry (2010 Q1) NAICS Industry Name Total Number Offices of Physicians (except Mental Health Specialists) 3, Family Planning Centers All Other Outpatient Care Centers Medical Laboratories All Other Miscellaneous Ambulatory Health Care Services General Medical and Surgical Hospitals Psychiatric and Substance Abuse Hospitals Specialty (except Psychiatric and Substance Abuse) Hospitals 68 Total 3,986 Source: U.S. Census County Business Patterns Division of Economic Development 23

28 Health Works Surveys that questions related to all three occupation groups of interest could be answered by these employers. Those categories are listed in Table 6.1. The employer survey included sections for each of the three types of employees, with instructions for employers to skip a category if they did not directly employ anyone in that category. Since the data included multiple locations for many companies, we instructed respondents to answer only with regard to their physical location when completing the survey. The following table shows the eight NAICS codes, descriptions, and the estimated number of establishments in the Louisiana, which was used as a basis for weighting the data to represent each sector. Table 6.2 shows the estimated number of offices for each of three broad industry groups and the estimated number of individuals directly employed in each of those industry groups. Offices of physicians make up the majority of physical locations and employ the largest number directly; however hospitals have the largest number of total employees. The all other category includes family planning centers, outpatient care centers, medical laboratories, and all other miscellaneous health care services. Figures 6.1 and 6.2 show the estimated proportion of locations and employees, respectively, by these three categories. The survey contained questions about staff as well as payments. In small offices, a single person may be able to answer both sets of questions. However, for large employers, some portions of the questionnaire might best be completed by a human resource manager while other portions might be more appropriate for someone involved in managing the establishment s finances. Therefore, we asked recipients to direct the survey to the person or persons who were most knowledgeable about the organization s human resources and finances. Because of the assistance provided by the Louisiana Hospital Association, we were able to address all hospital surveys directly to the hospital CEOs. The employer survey contains a section on each of the three broad occupation categories of interest. The survey asked some general questions for each occupational sub-category to assess the current workforce. In addition to the total number Table 6.2: Estimated Number of Offices and Employees by Industry Industry Type Estimated Number of Offices Offices of Physicians 3,410 68,340 Hospitals ,689 All Other 342 6,104 Estimated Number of Employees Figure 6.1: Estimated Number of Offices Figure 6.2: Estimated Number of Employees 24 LSU E. J. Ourso College of Business Administration

29 2012 Health Occupations Outlook of workers, they survey also included questions to assess the age and level of effort of workers to identify patterns in labor supply that might improve estimates of future availability of physicians. By examining existing data sources, we determined that hours varied considerably for physicians, while they tended to be more regular for clinical laboratory personnel and medical coders. Therefore, we asked the number of hours worked by physicians, but only how many part time and full time employees worked in the latter two categories. To provide a leading indicator of future employment, we also asked how many vacant positions exist for each occupation and how many of each the employer would like to employ in one year. Each of the three occupational sections also contained questions designed to address issues specific to each type of employee. For physicians in primary care specialties (family and general practitioners, general internists, obstetricians and gynecologists, pediatricians, and general surgeons), future supply of these physicians is a major concern. To better assess the current capacity for their services, the survey asks questions about accepting new patients and what types of payments would be accepted for new patients. In addition to assessing the overall availability of primary care physicians, the study considers whether shortages might arise for those covered by different types of insurance. To that end, the survey includes a series of questions designed to see capture how employers may respond to changes in the relative reimbursements for Medicaid, Medicare and private health insurance To accurately measure the number of workers in each of the clinical laboratory occupations, it is important consider how traditional employment statistics are organized. The Bureau of Labor Statistics (BLS) includes laboratory assistants in their counts of clinical laboratory technicians (Bureau of Labor Statistics, 2011a). However, because most employers would not group the two types of employees together, we divided clinical laboratory personnel into three categories: clinical laboratory technologists, clinical laboratory technicians, and laboratory assistants. For these employees, questions focused on the supply of clinical laboratory staff in Louisiana, with questions about the number of each type employed and how many were over 50 years old. Based on our review of existing data, we found that very few of these workers continued to be employed in these occupations for many years past age 50. Therefore, counts of the number of older workers provide an indicator of how many might be retired or near retirement by Beyond developing an accurate forecast of the number of clinical laboratory personnel, the questions of interest for this occupation group relate to the desired education of these workers by their employers. This information should help guide educational institutions in Louisiana in determining how many or what courses to offer. While not a complete picture of the benefits and costs of mandating higher levels of education, this information also could help inform future policy decisions. The medical coder questions made up the final section of the employer survey. In addition to the standard occupational inventory questions, we asked employers about the education and readiness among their employees for the upcoming switch to ICD-10. A major concern is that medical coders in Louisiana might not be prepared for the change that is coming in October of 2013, or that many medical coders might retire early if they are already near retirement. Finally, we asked questions about accountable care organizations, the patient-centered medical home, and the Bayou Health Network, which attempt to measure how widespread several new developments in health care are. Physician Survey The physician survey went to all Louisiana physicians with a practice area that focuses on primary care, which include general and family practitioners, general internists, pediatricians, obstetricians and gynecologists, and general surgeons. The sample was drawn from the Louisiana State Board of Medical Examiners list of physicians licensed by the state. From that list, we removed all physicians who practiced out-of-state, as well as all physicians with specialties outside of the primary care areas. Physicians with no specialty listed in the licensing data were included in the survey in order to capture their specialties and provide the most comprehensive measure of physicians possible. Physicians listing multiple specialties were included if at least one of their specialties fit into one of the primary care categories. Based on the practice area from the licensing database and the responses from the survey, this approach provides the best information possible for estimating the total number of primary care physicians. The physician survey included basic education and occupational inventory questions, such as asking for the physician s medical school, practice area and the number of hours worked. In addition, age and several questions related to future labor supply provided helpful inputs to the forecasting component of the study. As in the employer survey, the physician survey includes questions about patient care and types of insurance accepted to help better gauge the availability of care to subgroups of the population. Finally, a section of employer questions was also included to ensure that information regarding health employers was obtained in cases where small practices might receive both surveys and return only one. We received responses from 1,133 physicians in our survey population. After accounting for bad addresses and disconnected phone numbers the response rate for the physician survey was calculated to be 18.3 percent. About 15 percent of the physicians surveyed had a blank specialty in the licensing data, so a number of responding physicians indicated that their specialty was something other than one of the five primary care specialties. Because the survey focused on Division of Economic Development 25

30 Health Works Surveys Figure 6.3: Age Distribution of Physicians in Louisiana primary care, those physicians who received surveys but were not primary care physicians were slightly less likely to complete the survey. Considering only physicians with primary care specialties, the response rate is a bit higher at 18.6 percent. To ensure that the sample best represents all physicians, weights were created using practice area and five-year age groups, the two characteristics for which good population estimates are available. The licensing database did not contain an age variable, so physician age was imputed using the assumption that physicians were approximately 27 years old upon graduation from medical school. For physicians with missing medical school graduation dates, we used a hot deck method to impute graduation year based on physicians date of first license issue and medical school location under the assumption that students from in state schools are more likely to obtain a license immediately while those from out of state are more likely to obtain a license at a later point. Then, base weights were created for all survey respondents based on age and practice area variables, so that physicians who responded would represent all Louisiana primary care physicians in their five-year age range and practice area groups. With a goal of estimating the number of working physicians, an adjustment was made to the weights to account for retired physicians that continue to renew their licenses and thus may be overrepresented by simply weighting to the licensing database. Thus, we used the age distribution of the American Community Survey (Census Bureau, 2012a) for the years 2008 to 2010 (shown in blue) to adjust the weighting so that older physicians have a somewhat lower weight. Primary Care Physician Results While questions about physicians were asked of employers, this section focuses on responses to the physician survey assuming that those direct responses best reflect the entire population of physicians working in Louisiana in one of the five primary care practice areas. However, in some cases, responses from employers provided interesting context for the physician responses and are included where appropriate. Physician Inventory After weighting to ensure the sample best represents all physicians, the estimated numbers of physicians in each of the five physician categories are shown in Table 6.4. The difference between licensed physicians and active physicians accounts for those physicians who are retired or not practicing for any other reason, but maintain a Louisiana license. The five 26 LSU E. J. Ourso College of Business Administration

31 2012 Health Occupations Outlook physician specialties shown in Table 6.3 were selected as the focus of the study because of their expected role in providing primary care services to Louisiana s population. However, to provide additional insight into the kinds of services physicians provide, the survey also asked physicians whether they considered themselves to be providing primary care services. The vast majority of family and general practitioners as well as obstetricians and gynecologists are providing primary care services, while less than two-thirds of internal medicine physicians provide those services as a primary function of their practice. It is also worth noting that a small number of specialists indicated that they provide primary care services. However, we expect that those physicians do not provide primary care services as the main function of their practice and thus were not added to the scope of the project. Considering these responses, of the 5,923 physicians working in the five categories of interest to the study, the best estimate of the number of primary care physicians is actually 4,570. While access to health care is important to this study, the primary focus is forecasting the future demand for physicians. Therefore, to ensure consistency with the broader employment forecast, the total number of active physicians practicing in each of the five categories below is used as the primary measure of physicians in primary care specialties for most of the discussion that follows. The number of primary care physicians was also estimated using responses from the employer survey. Considering all physicians working at the employer locations, the estimated number of primary care is nearly twice as high as the estimate from the physician survey and licensing data. This is caused by the fact that many physicians work at multiple locations (e.g. a physician with a private practice who also visits patients in the hospital). While self-employed physicians satisfy an important portion of the overall demand for physician services, responses from employers provide helpful information for understanding the demand for additional physicians. Table 6.4 shows the results of a series of questions regarding vacant positions and how many physicians the employer would like to employ in one year s time. The vacancy rate represents the number of open positions employers are looking to fill divided by the total number of physicians in each practice area. This measure shows current unmet demand. By comparison, the average vacancy rate across all occupations from the most recent Louisiana Job Vacancy Survey was 1.1 percent, down from a post-katrina high of 6.9 percent (Louisiana Workforce Commission, 2011). This historical context demonstrates clearly a significant unmet demand for physicians in today s labor market. Also shown in Table 6.4 is the level of growth employers see coming over the next year. The 9.9 percent average growth rate across all physician groups again represents a very robust outlook for demand in the coming year with employment statewide projected to grow by only 0.8 percent according to the most recent Louisiana Economic Outlook (Scott & Richardson, 2011). To better understand the supply of physicians, the survey included questions about the number of hours currently worked as well as how many hours each physician planned to work in one, five, and ten years. The estimated average weekly hours are shown in Table 6.5. On average, today s physicians anticipate working fewer hours in the future than they do currently, with average hours worked by primary care physicians going from 50.5 this year to 43.2 in This is driven by the common practice of physicians cutting back on their hours as they age. Of course, this trend of decreasing hours does not represent expected changes in hours of all physicians in the future because new physicians will begin working and most likely commit to longer hours as is common among physicians early in a career. Average weekly hours by age are shown in Table 6.6, which reflect variations in labor supply throughout a career. The number of physicians and average hours exclude physicians with a license but who are not working. Overall, primary care physicians in Louisiana work a little over 50 per week on average until age 65, when the physicians who continue to work tend to work fewer hours. Figure 6.4 shows the portion of physicians that will stop practicing in Louisiana within the next ten years, which accounts for retirement and those who may be planning to Table 6.3: Estimated Number of Physicians and Primary Care Activities Practice Area Licensed Physicians Active Physicians Percent Primary Care Family/ General Practice 1,502 1, % Internal Medicine 2,435 2, % Obstetrics and Gynecology % Pediatrics 1,144 1, % General Surgery % Total 6,463 5, % Division of Economic Development 27

32 Health Works Surveys move to another state. Among the 5,923 active physicians in the five primary care practice areas, 29 percent plan to stop working in the state by This high portion of physicians leaving their practice in the state will contribute to even greater demand for new physicians than has been typical of recent years. To better understand how these retirements will impact the supply of labor in the forecast year of 2020, physicians were asked a follow up question about when they might retire. Based on this information, a detailed breakdown of changes in practice plans by specialty for the year 2020 are shown in Table 6.7. Physician Payment Types To better understand physician availability, several questions were included to assess physician supply for specific payment types. Physician payments received by source according to both the employer and physician surveys are shown in Table 6.8. The six payment types considered are private insurance; Medicare; Medicaid (including LaCHIP); TRICARE, CHAMPUS/ CHAMP-VA, VA, or other military; self-pay; and other. Because some employers responded for a group of physicians, the averages from the two surveys are not equal. The results from the physician survey provide a more direct measure of access for individuals. To examine access more closely, responses were summarized to determine acceptance of new patients for each payment type. Figure 6.5 shows the percent of physicians accepting new patients by payment type according to the physician survey. As expected, private insurance is the most widely accepted type of payment for new primary care patients in Louisiana, with about 85% of physicians accepting new privately insured patients and Medicaid is the least widely accepted, with about 54% of physicians accepting new Medicaid patients. Table 6.4: Employer Demands for Primary Care Physicians by Practice Area Practice Area Vacancy Rate 1-Year Growth Family/General Practitioners 8.3% 6.4% General Internists 12.9% 22.7% Obstetricians and Gynecologists 2.8% 9.0% Pediatricians 5.1% 6.1% General Surgeons 6.1% 5.9% Total 7.8% 9.9% Table 6.5: Estimated Average Weekly Physician Hours Year Average Hours This Year 50.5 Next Year 49.1 In 5 Years 46.2 In 10 Years LSU E. J. Ourso College of Business Administration

33 2012 Health Occupations Outlook Table 6.6: Estimated Weekly Physician Hours by Age Age Number of Physicians Average Hours Total 5, Figure 6.4: Primary Care Physicians Not Practicing in 10 Years Table 6.7: Current Physicians Not Practicing in 2020 by Specialty Practice Type Percent Total Family/ General Practice 22.3% 307 Internal Medicine 22.1% 487 Obstetrics and Gynecology 31.9% 233 Pediatrics 24.1% 256 General Surgery 24.8% 136 Total 24.3% 1,419 Division of Economic Development 29

34 Health Works Surveys Table 6.8: Survey Results for Current Physician Effort by Payment Type Payment Type Insurance Medicare Medicaid Military Self-Pay Other Physician Survey 42.3% 24.5% 20.5% 3.9% 7.2% 1.6% Employer survey 33.4% 24.0% 23.8% 4.2% 10.3% 4.3% Figure 6.5: Estimated Acceptance of New Patients by Payment Type Figure 6.6 shows the results of the same question asked of employers, with the results shown by industry. The most interesting result is that hospitals are more likely to accept all types of payment, while offices of physicians show a pattern very similar to what physicians reported on the physician survey. As shown in Figure 6.5, the two payment types with most limited access are Medicare and Medicaid. Table 6.9 shows the percentage of employers accepting each of these two types of payments according to the employer survey. Hospitals are more likely to accept Medicare and Medicaid, and almost all hospitals accept Medicare. Another perspective on access to care for Medicare and Medicaid patients is provided in Table 6.10, which shows the percent of physicians who do not accept each type of payment by specialty. The most striking results are the high percent of family and general practitioners and general internists not accepting Medicaid. A very high percentage of pediatricians do not accept Medicare, but this is expected. To better anticipate the availability of physicians to those covered by Medicare and Medicaid, the data was broken out by physician age. Figure 6.7 shows that older physicians are less likely to accept Medicare and slightly more likely to accept Medicaid than are younger physicians. Setting aside potential issues with regard to overall physician supply, this result suggests that it may be more difficult for Medicaid patients to find care in the future when those older physicians have retired. However, the cross sectional nature of this survey makes it impossible to determine if this is a cohort effect or a lifecycle effect (i.e. is the older generation of physicians simply more inclined to treat those with Medicaid relative to Medicare or do physicians tend to shift away from Medicare and toward Medicaid later in their careers). 30 LSU E. J. Ourso College of Business Administration

35 2012 Health Occupations Outlook Figure 6.6: Industry Comparison of New Patients by Payment Type Table 6.9: Estimated Access to Medicare and Medicaid by Industry Industry Type % Not Accepting Medicare Patients % Not Accepting Medicaid Patients Offices of Physicians 27.4% 39.7% Hospitals 2.7% 21.5% All Other 35.2% 37.6% Table 6.10: Physicians Not Accepting Medicare and Medicaid by Practice Type Practice Area Medicare Medicaid Family and General Practitioners 11.5% 38.1% General Internists 15.8% 43.2% Obstetricians and Gynecologists 13.6% 20.9% Pediatricians 87.7% 20.1% General Surgeons 11.6% 22.8% Total 26.1% 31.8% Division of Economic Development 31

36 Health Works Surveys Figure 6.7: Estimated Percent of Physicians Not Accepting Medicare and Medicaid by Physician Age Changes in Payment Types We asked both physicians and employers how their current distribution of payment types for primary care services would change under various conditions. The remainder of this section focuses on how payment types would change relative to the current distribution of payments shown in Table 6.9 under each scenario. First, respondents were asked about the distribution of payments in 2006, a time when the relative payments for primary care services to overall Medicare payments were lower than today. Holding constant other changes to the health care market, this change would be expected to result in a greater share of physician services directed toward Medicare in 2011 than in 2006, or conversely a drop in Medicare, in Figures 6.8a and 6.8b. However, survey results showed no significant change in the mix of patients suggesting that other factors in the health care market masked the response of physicians to this change in reimbursement (or that the response to changes in reimbursement was relatively small). The second condition we considered that might impact the mix of patients was a reduction in Medicare payments by 20 percent, which is similar to what would happen if the sustainable growth rate were implemented and no doc fix were passed. Figures 6.9a and 6.9b show similar responses by physicians and employers, with Medicare patients being dropped in favor of the privately insured and out-out-pocket payers. It is interesting that both groups report reducing their Medicaid patients when only Medicare reimbursements decrease. One explanation is that there are fixed costs associated with dealing with the two government health insurance programs, and that some physicians would react to Medicaid reimbursement reductions by eliminating all Medicare and Medicaid patients. The third condition we considered that might impact the mix of patients was an increase in Medicaid reimbursements to equal Medicare reimbursements, which is a temporary change mandated by the ACA. Figures 6.10a and 6.10b show the results from physicians and employers, respectively. In Louisiana, the per-visit change is relatively small since Medicaid is currently 90 percent of Medicare levels (Zucherman, Williams, & Stockley, 2009). Perhaps because of the relatively small impact on per-visit reimbursements, the survey results show a fairly small impact on the proportion of Medicaid patients and a surprisingly larger decrease in the proportion of Medicare patients. The survey results do not provide a clear explanation 32 LSU E. J. Ourso College of Business Administration

37 2012 Health Occupations Outlook Figure 6.8a: Physician Payment Distribution in 2006 Relative to 2011 Figure 6.8b: Employer Payment Distribution in 2006 Relative to 2011 for why physicians and employers both would expect to reduce Medicare payments if Medicaid payments were increased to equal Medicare, but perhaps physicians expect that such a link would ultimately lead to lower Medicare payment growth down the road. Lastly, we asked both physicians and employers how they would react if Medicaid reimbursements decreased by 20 percent. The results for that condition are in Figures 6.11a and 6.11b. Both groups reported decreasing their percentage of patients who have Medicaid coverage by almost 10 percent. Physician Training and Recruitment To provide a better understanding of new physicians entering the market, physicians were asked for the location of the medical school from which they graduated. Almost two-thirds of today s Louisiana physicians graduated from a school in Louisiana. About Division of Economic Development 33

38 Health Works Surveys Figure 6.9a: Physician Reaction if Medicare Reduced by 20% Figure 6.9b: Employer Reaction if Medicare Reduced by 20% ten percent of the physicians graduated from medical schools in foreign countries. Of other U.S. states, Texas was the most common, but small percentages of Louisiana physicians attended medical schools in a wide variety of other states. Additionally, we asked physicians where they completed their residency. Since physicians in the U.S. are required to complete a residency in the country, no physicians did a residency in a foreign country. About two-thirds of Louisiana s active physicians completed a residency in Louisiana. New York was the next most popular state, possibly due to it being a popular location for foreign physicians to go for their American residency. Texas was the third-most popular state, and small numbers of physicians came to Louisiana after completing residency programs in a variety of other states. The relationship between medical school location and residency location is shown in Table The rows show the percent of physicians by medical school location while the columns show percent of physicians by residency location. Each training 34 LSU E. J. Ourso College of Business Administration

39 2012 Health Occupations Outlook Figure 6.10a: Physician Reaction if Medicaid Increased to Medicare Figure 6.10b: Employer Reaction if Medicaid Increased to Medicare location is grouped into the categories of Louisiana, the rest of the South region (as defined by the Census Bureau), the rest of the United States, and foreign country for medical schools only. Over half of primary care physicians in Louisiana attended one of Louisiana s medical schools and completed a residency in Louisiana, while more than three-fourths did at least one of the two in Louisiana. primary care physicians across all industries. It should be noted that a significant number of employers did not respond to the question, so the summary in Figure 6.12 represents the percentage of employers utilizing each method out of the set of employers responding to the question. The most popular methods for recruiting primary care physicians are online, relationships with medical schools, and employment agencies. The employer survey also asked how physicians are recruited. Figure 6.12 shows the recruitment methods used to recruit Division of Economic Development 35

40 Health Works Surveys Figure 6.11a: Physician Reaction if Medicaid Payments Reduced by 20% Figure 6.11b: Employer Reaction if Medicaid Payments Reduced by 20% Open-Ended Physician Responses With the opportunity to hear directly from physicians regarding some of the upcoming changes to the health care sector, the survey included a couple of open-ended questions. Table 6.12 shows the results of an open-ended question asking physicians what they would do if their malpractice insurance premiums increased by 20 percent. We grouped the responses into several major categories. Many physicians said they would react in a specific way, such as retiring or refusing to accept Medicaid patients. Because many physicians were not inclined to provide a response, this table lists the percent of responses that fell into in each category out of the total number of physicians responding to the question. While an alarmingly high percentage of physicians claimed that they would retire, approximately 29 percent of physicians said that they were planning to stop practicing within the next 36 LSU E. J. Ourso College of Business Administration

41 2012 Health Occupations Outlook Table 6.11: Louisiana Physicians by Location of Medical School and Residency Residency Medical School Louisiana Other Southern States Other U.S. Regions Total: Louisiana 52.1% 9.2% 5.3% 66.6% Other Southern States 4.7% 6.5% 2.2% 13.4% Other U.S. Regions 2.7% 0.5% 6.8% 10.1% Foreign Country 3.4% 0.6% 5.9% 9.9% Total: 62.9% 16.8% 20.2% 100.0% Figure 6.12: Employers Methods for Recruiting Primary Care Physicians ten years as shown in Figure 6.4. It is likely that many of those indicating they would retire are already close to retirement and such a change would potentially accelerate those plans. It is also likely that the majority of physicians who declined to answer this question would have weaker reactions to this subject. Table 6.13 summarizes the responses to a question asking physicians how they have responded to the Affordable Care Act. The most common response group included any comment indicating that the physician would essentially continue to do the same things they had been doing. Many physician responses essentially stated why they did not like the ACA without offering any comments about changing their practice, which we categorized as Dislikes ACA, but no reported change. Similar to the previous question, the percentages are only out of the total number of physicians who responded. Division of Economic Development 37

42 Health Works Surveys Table 6.12: Common Responses to Malpractice Insurance Increase of 20 Percent Response I would retire 22.6% I would not do anything differently 19.3% I would have to make a major change to my practice 18.5% This change would not affect me 15.4% Dislikes change, but no reported impact on practice 10.2% I would have to change the type of patients I see (ex: Medicaid) 8.1% Table 6.13: Common Reactions to the Affordable Care Act Response I would not do anything differently 29.1% Dislikes ACA, but no reported change 14.2% Implement Electronic Medical Records 8.0% This change would not affect me 7.7% I would have to change the type of patients I see (ex: Medicaid) 5.2% I would have to make a major change to my practice 3.8% I would retire 3.1% Technologists and Technicians Results Technologist and Technician Inventory The total number of clinical laboratory personnel employed at medical laboratories, hospitals, and other medical establishments included in the survey is shown in Table 6.14, along with the percentage of clinical laboratory personnel employed directly. To avoid double counting, employers were asked to answer subsequent questions only with regard to directly employed workers. These estimates are slightly lower than OES estimates; however, the discrepancy is easily explained by the relatively small number of health employers not included in the sampling frame. While we expected some laboratory personnel to work in hospitals and other large medical organizations, the surprising result is the relatively small number of clinical laboratory personnel employed by medical laboratories. This indicates a likely pattern that much of the lab work done for Louisiana residents outside of hospitals is sent to larger regional laboratories out of state. Table 6.15 shows the results of four questions we asked employers regarding each of three types of clinical laboratory personnel: clinical laboratory technologists, clinical laboratory technicians, and laboratory assistants. For each type, we asked how many were employed at the organization full-time, how many were employed part-time, how many vacant positions for each type existed, and how many of each type the employer would like to employ in one year s time. In the context of historical vacancy rates, these results show strong demand for all three types of employees, especially clinical laboratory assistants. The results for 1-year growth are mixed, possibly indicating a shift in the mix of workers needed by employers. However, the results may be slightly biased downward because some employers appeared to respond as if the question were intended to capture only new jobs rather than total jobs in one year. Ultimately, the growth results should be interpreted as showing stronger relative growth for technicians than for the other two categories. Since there is no survey of individual clinical laboratory personnel, employers were asked for information on the ages of clinical laboratory personnel to assess the potential impact of retirement on labor supply in the coming years. Table 6.16 shows the percentage of the total number of each type of clinical laboratory worker in each of the age ranges from the survey, as well as the percentage that are less than 50 years old. Technologists tend to 38 LSU E. J. Ourso College of Business Administration

43 2012 Health Occupations Outlook Table 6.14: Estimated Lab Personnel by Industry Industry Type Total Lab Personnel Percent Directly Employed Medical Laboratories % Hospitals 2, % All Other 1, % Total 4, % Table 6.15: Estimated Types of Clinical Laboratory Personnel Occupation Full-Time Part-Time Vacancy Rate 1-Year Growth Technologists 2, % -4.5% Technicians % 6.8% Assistants % -3.2% Table 6.16: Estimated Ages of Clinical Laboratory Personnel Occupation Under to to or Older Technologists 64.7% 15.0% 9.0% 11.4% Technicians 75.7% 11.1% 6.8% 6.4% Assistants 71.6% 18.4% 5.7% 4.3% be older than technicians and assistants, which may somewhat be explained by the fact it is a more senior-level position. Both because of their greater numbers and their higher average age, clinical laboratory technologists will require more emphasis in replacing their numbers than technicians and assistants, but training for all three categories is still important to provide employers with the right mix of workers. Technologist and Technician Education Employers were asked about their preferred level of education for clinical laboratory personnel. Figures 6.13 and 6.14 show both the current and desired educational levels for technologists and technicians by industry group among employers of each group of workers. While nearly 75 percent of hospital-employed technologists have a bachelor s degree, hospitals would prefer that 90 percent have a bachelor s degree. In medical laboratories, over 60 percent of technologists currently have a bachelor s degree. However, medical laboratories indicated they would prefer slightly fewer technologists to have a bachelor s degree. As for laboratory technicians, nearly 70 percent of hospitals employed technicians have an associate s degree, while hospitals would prefer slightly more to have an associate s degree. For medical laboratories, nearly 60 percent of technicians have an associate s degree, but those employers would prefer a workforce with slightly fewer associate s degrees, presumably as a way to reduce costs. While there are significant differences in the proportion of degree holding in different industry settings, there is a strong pattern across these results showing that employers want a laboratory workforce that is about as educated as the one they currently have. These results make it clear that while there is certainly a need for more clinical laboratory technologists and technicians in Louisiana, current clinical laboratory personnel education levels are consistent with employer needs and employers are not demanding all technicians to have an associate s degree or all technologists to have a bachelor s degree. Even though having a bachelor s degree is important for clinical laboratory technologists who wish to work for a hospital, there is room in the clinical laboratory workforce for technologists with and without bachelor s degrees and technicians with and without associate s degrees. Division of Economic Development 39

44 Health Works Surveys Figure 6.13: Estimated Actual and Preferred Education of Technologists Medical Coders Results Medical Coder Inventory Table 6.17 shows the results of the inventory questions we asked employers regarding medical coders. We asked how many were directly employed full-time and part-time, how many vacant positions they had, and how many the employer would like to employ in one year s time. All of these numbers are aggregate totals across all employers who responded to the survey. The vast majority of medical coders work in offices of physicians. The high vacancy rates and high rate of growth reported by employers in the coming year indicate a very strong demand for new coders. In fact, the kind of growth reported by employers is unlikely to be met without significant recruiting from out of state or a major change in training locally. Additionally, we asked employers how many medical coders are in three age categories that are close to possible retirement age. Table 6.18 shows the percentage of the total number of medical coders in each age range. Coders in offices of physicians tend to be slightly older than other coders, a pattern that explains some of the demand for new workers in that industry. Almost 40% of physician office coders who are over 50 years old will need replacements with the next decade or two. ICD-10 Finally, the employer survey included a number of questions about medical coders and ICD-10 implementation (see the section on ICD-10 for more details). Table 6.19 shows percentages of medical coders who have taken courses in Anatomy, Physiology, and Medical Terminology, the percent with plans to learn ICD-10 before the October 1st, 2013 deadline, and the percent retiring instead of learning ICD- 10. Hospital coders are ahead of other coders in taking the necessary coursework, and more of them have plans to learn ICD-10. The percentage of coders employers expect to retire specifically to specifically avoid the ICD-10 implementation is relatively small, but the implementation of ICD-10 will likely coincide with an increase in demand for ICD-10 capable coders. 40 LSU E. J. Ourso College of Business Administration

45 2012 Health Occupations Outlook Figure 6.14: Estimated Actual and Preferred Education of Technicians General Results Several questions were included in the surveys to collect data to provide background information on Louisiana s health care industry. This context informs the outlook for all three sets of occupations and is summarized below. Patient Capacity Employers and physicians were asked about their organization s ability to meet the needs of its patients. Figure 6.15 shows responses by employers for offices of physicians, hospitals, and other employers who report having excess, adequate, and insufficient capacity to handle their patient load. Most report adequate capacity, but hospitals are somewhat more likely to report insufficient capacity and less likely to report excess capacity. Overall, about 70 percent of physicians reported adequate capacity with the remaining 30 percent split evenly between those reporting excess capacity and those reporting insufficient capacity. Figure 6.16 shows responses of physicians by practice area. One interesting finding is that general surgeons and internists are much more likely to report having insufficient capacity than the other three types of primary care physicians. Changes in the Healthcare Industry Figure 6.17a shows how many employers said they used electronic medical records (EMR), as well as how many did not. A slight majority of health employers in Louisiana report already using electronic medical records. Figure 6.17b shows how many physicians said they used electronic medical records as well as how many did not. Almost three-fourths of physicians now use EMR. The best explanation for why physicians use EMR at a higher rate than employers is that larger establishments are more likely to use EMR and physicians who practice in more than one location may report using medical records even if it applies only to one setting (e.g. a hospital). Also, these results include employers who do not employ any physicians, such as laboratories. Establishments reporting no EMR still have a few years until the CMS starts to impose penalties for not using EMR, if the rules are applicable to them. Division of Economic Development 41

46 Health Works Surveys Table 6.17: Estimated Medical Coders by Industry Industry Type Full-Time Part-Time Vacancy Rate 1-Year Growth Offices of Physicians 2, % 25.1% Hospitals % 9.3% All Other % 3.3% Total 3, % 21.0% Table 6.18: Estimated Medical Coder Ages by Industry Industry Type Under to to or Older Offices of Physicians 60.8% 21.6% 9.6% 8.0% Hospitals 67.3% 14.8% 8.8% 9.1% All Other 69.6% 9.9% 8.6% 11.8% Total 68.1% 14.7% 8.0% 9.3% Table 6.19: Estimated Medical Coder Skills by Industry Industry Type Taken Anatomy Taken Physiology Taken Medical Terminology Plan to Learn ICD-10 Retiring Instead of Learning ICD-10 Offices of Physicians 44.2% 35.8% 65.2% 85.9% 5.9% Hospitals 73.8% 72.8% 76.7% 94.8% 4.9% All Other 45.5% 42.0% 68.7% 89.3% 11.4% Total 50.4% 43.8% 67.7% 87.9% 5.9% A related change is a shift toward electronic billing systems (EBS), which is already required for processing Medicare and Medicaid payments. While a shift to an electronic billing system comes with some fixed cost, greater utilization of technology (especially if it is linked to EMR) can also improve efficiency in billing. Figure 6.18a shows the percent of employers using EBS while Figure 6.18b shows the results of a similar question asked of physicians. Another general trend that was assessed in the surveys is the increased prevalence of patient-centered medical homes. Table 6.20 shows the proportion of employers who are or have plans to become a patient-centered medical home by industry. Since many of these organizations are places that inherently would not fit the definition of a medical home, such as a laboratory, a not applicable option was included. Approximately 15 percent of employers are patient-centered medical homes (or contain one) across industries, and there does not seem to be a large amount of variation between industries. The only exception is a relatively large number of hospitals that are planning to create a patient-centered medical home. The results of the question asking physicians if they practice in a medical home are shown in Figure Even though relatively few physicians said they currently practice in a medical home, almost as many physicians are planning to join or become a medical home indicating the continuing trend of increased interest in this structure for providing care. 42 LSU E. J. Ourso College of Business Administration

47 2012 Health Occupations Outlook Figure 6.15: Estimated Capacity to Provide for Patients by Industry Figure 6.16: Estimated Capacity to Provide for Patients by Area Division of Economic Development 43

48 Health Works Surveys Figure 6.17a: Estimated Employers Use of Electronic Medical Records Figure 6.17b: Estimated Physician Use of Electronic Medical Records Results by physician practice area are provided in Table The patient-centered medical home is most popular with pediatricians, which makes sense given that the idea originated from the American Academy of Pediatrics. Obstetricians and Gynecologists and General Surgeons are the least likely to form a medical home. Another new development in the provision of health care is the increasing organization of providers into accountable care organizations (ACO). Table 6.22 shows employer participation and interest in ACOs, with results broken out by industry. About 14 percent of offices of physicians are part of an ACO and 3 percent are planning to join one. Hospitals show greater interest in the idea, with nearly 11 percent of hospitals currently part of an ACO but 26 percent having plans to join one. Figure 6.20 shows the results of asking physicians if they are part of an ACO. Since this is a new concept, it is not surprising that a relatively high proportion (27.8 percent) said that they were unsure. These responses indicate that ACOs are still new 44 LSU E. J. Ourso College of Business Administration

49 2012 Health Occupations Outlook Figure 6.18a: Estimated Employer Use of Electronic Billing Systems Figure 6.18b: Estimated Physician Use of Electronic Billing Systems Division of Economic Development 45

50 Health Works Surveys Table 6.20: Estimated Patient-Centered Medical Home Prevalence by Industry Industry Type Yes No Planning To Join Unsure Not Applicable Offices of Physicians 14.3% 58.3% 6.0% 14.3% 7.1% Hospitals 16.0% 52.8% 18.7% 5.3% 7.2% All Other 15.4% 60.5% 0.0% 5.1% 19.0% Total 14.4% 58.1% 6.4% 13.5% 7.6% Figure 6.19: Estimated Patient-Centered Medical Home Prevalence Table 6.21: Estimated Patient-Centered Medical Home Prevalence by Practice Type Practice Type Yes No Planning To Join Unsure Not Applicable Family/ General Practice 17.0% 43.4% 17.4% 17.4% 4.7% Internal Medicine 25.7% 36.3% 13.1% 15.5% 9.4% Obstetrics and Gynecology 9.7% 60.1% 10.6% 11.3% 8.2% Pediatrics 35.4% 24.4% 15.2% 17.1% 7.8% General Surgery 8.5% 60.7% 16.9% 9.3% 4.5% Total 21.2% 41.9% 14.6% 15.1% 7.2% 46 LSU E. J. Ourso College of Business Administration

51 2012 Health Occupations Outlook Table 6.22: Estimated Accountable Care Organization Prevalence by Industry Industry Type Yes No Planning to Join Unsure Offices of Physicians 14.1% 62.9% 3.1% 19.9% Hospitals 10.6% 58.3% 26.0% 5.2% All Other 16.7% 39.0% 0.0% 44.4% Total 14.0% 61.6% 4.2% 20.2% Figure 6.20: Physicians in Accountable Care Organizations Part of ACO to Louisiana but that their prevalence is growing, with almost half as many physicians as currently practice in an ACO having plans to join one. The results from the physician survey are consistent with those from the employer survey on this issue and differences between specialties were insignificant. The Louisiana Medicaid program has recently taken a major step toward changing the way care is delivered by partially privatizing the program with the Bayou Health Network. We asked health employers if they were a part of a Bayou Health coordinated care network. A summary of employer responses are provided in Figure 6.21 with overall participation indicated in the larger shaded boxes. These percentages do not add up to 100 percent because employers can be affiliated with multiple insurers. Further, because employers may join one or more network without joining all five, the total percent of employers in each industry participating in any Bayou Health network is higher than the percent for any one network (e.g percent of offices of physicians participate in one of the five networks while just under fifty percent of physicians participate in United Healthcare Community Plan, the most common plan). Hospitals were much more likely than offices of physicians and other employers to be part of Bayou Health. Each of the five insurers has roughly equal popularity. The physician survey also included questions about the Bayou Health Network. Figure 6.22 shows the percentage of physicians who said they were affiliated with at least one Bayou Health insurer. Given that 31.1 percent of physicians do not currently participate in Medicaid, this result shows that some physicians are late in signing up with at least one Bayou Health insurer, or that some physicians may be no longer be serving Division of Economic Development 47

52 Health Works Surveys Figure 6.21: Estimate Bayou Health Network Affiliation by Industry Figure 6.22: Estimated Percent of Physicians Participating in the Bayou Health Network 48 LSU E. J. Ourso College of Business Administration