The 2004 election saw President George W. Bush (Republican) reelected,

Transcription

1 Voters And Health Care In The 2004 Election Health care concerns were not foremost on voters minds when they went to the polls in November by Robert J. Blendon, Mollyann Brodie, Drew E. Altman, John M. Benson, and Elizabeth C. Hamel ABSTRACT: Exit polls showed that health care was a second-tier issue in the 2004 presidential race and that it was more important to Democratic than Republican voters. Those who considered health care the most important issue in their voting decision voted overwhelmingly for John Kerry. An analysis of national opinion surveys and exit polls shows that Republican and Democratic voters expressed dramatically differing views on key health policy issues. With the exception of malpractice reform and reducing federal health spending, health care is not likely to be a top issue for the Bush administration s second term. The 2004 election saw President George W. Bush (Republican) reelected, winning the popular vote by a three-million-vote margin over Sen. John Kerry (Democrat) and a majority of electoral votes. The Republicans gained four seats in the Senate, where they now hold a advantage, with one Independent. 1 The Republicans gained three seats in the U.S. House of Representatives, and they now have a advantage, with one Independent. 2 What role, if any, did health care play in the 2004 presidential election? Elections are important in policy areas like health care, for two main reasons. If an issue ranks high among voters, it often becomes a priority for the new administration and Congress. Also, if one party holds both the presidency and a majority of seats in Congress, the views of voters who identify with that party help predict the direction of the legislative agenda. 3 Because Republicans hold both the presidency and a majority in both houses of Congress, the views of their supporters are likely to have a stronger influence on the direction of health care policy than the views of the Democrats supporters. In this paper we examine fourteen national surveys and exit polls, taken before Bob Blendon (rblendon@hsph.harvard.edu) is a professor of health policy and political analysis at the Harvard School of Public Health in Boston and the John F. Kennedy School of Government in Cambridge, Massachusetts. Mollyann Brodie is vice president and director of public opinion and media research, at the Henry J. Kaiser Family Foundation, Menlo Park, California. Drew Altman is the foundation s president. John Benson is managing director of the Harvard Opinion Research Program, Harvard School of Public Health. Elizabeth Hamel is associate director of public opinion and media research, Kaiser Family Foundation. W March 2005 DOI /hlthaff.W Project HOPE The People-to-People Health Foundation, Inc.

2 Health & Election and after the 2004 election, to determine, if possible, the views of the electorate regarding health care and its importance on the national agenda. Data Sources And Methods The survey data in this paper are derived from three main sources. The first is a Henry J. Kaiser Family Foundation/Harvard School of Public Health pre-election survey, conducted by telephone October 2004 with a nationally representative sample of 1,202 adults, including 1,064 registered voters. 4 The second source consists of ten other national surveys conducted between January and December Each survey included 800 1,833 randomly selected registered or likely voters, or self-described voters, in the 2004 election. 5 The third consists of three exit polls conducted on Election Day, 2 November These include a national exit poll of 5,154 voters and a California exit poll of 3,357 voters conducted by the Los Angeles Times, and a national exit poll of 13,600 voters conducted for the National Election Pool, a consortium that provides exit polling data for the main television networks and a number of major newspapers, by Edison Media Research and Mitofsky International. In exit polls, questionnaires are distributed at voting places that have been selected to obtain a sample that will reflect the opinions of voters throughout the country or state. 6 Study Results Theroleofissues.As one examines the role of health care in the 2004 presidential election, it is important to keep in mind that the candidates positions on issues such as health care are only one of the criteria voters use to decide how they vote. Two surveys, just before and just after the 2004 election, showed that the characteristics of the candidates were more important than their stands on issues to at least a plurality of voters. In the postelection survey, a majority of self-described voters said that the candidates leadership skills, character, values, and experience (53 percent) were more important than where the candidates stood on issues (39 percent). 7 When voters were asked in the national exit polls about the most important issue determining their vote, moral values ranked first in both the National Election Pool and Los Angeles Times exit polls, followed by economy/jobs, terrorism, and Iraq. Following the election, there was much controversy over the meaning of the term moral values as the issue that ranked first in both polls. Two recent studies have shown that voters meant different things when they said that moral values were an important issue. While some reported voting based on specific policy issues, such as abortion or gay marriage, others said that they were concerned about the character, religious values, and leadership abilities of the candidates themselves. 8 Health care as an election issue. Health care was a second-tier issue in the 2004 presidential election, ranking fifth among voters as the most important issue in the National Election Pool exit poll. In the Los Angeles Times national exit poll, which HEALTH AFFAIRS ~ Web Exclusive W5-87

3 included a longer list of issues, health care was tied for seventh place, and Medicare/prescription drugs ranked eleventh. Health care ranked behind moral values, economy/jobs, terrorism, and Iraq as a voting issue in both exit polls. 9 Health care declined as a potential voting issue over the course of the year. In a January 2004 poll, registered voters ranked health care and Medicare second as a voting issue (19 percent), behind only the economy and jobs (36 percent) and ahead of terrorism (14 percent) and Iraq (11 percent). 10 In each national exit poll, health care was cited as the most important issue by many more Kerry than Bush voters (Exhibit 1). 11 About three-fourths (77 percent) of voters who said that health care was the most important issue voted for Kerry, while only 23 percent voted for Bush. 12 Healthcarevotersweremorelikelythanothervoterstobewomen,agesixtyfive and older, and African Americans, andtohavehouseholdincomesoflessthan $50,000 per year. Sixty percent of health care voters were Democrats, 17 percent were Republicans, and 23 percent were Independents. 13 In a postelection survey, a majority of uninsured voters said that they had voted EXHIBIT 1 Most Important Issues In Deciding The 2004 Presidential Vote, Among Voters Nationwide National Election Pool exit poll a Total voters Bush voters Kerry voters 1. Moral values (22%) 2. Economy/jobs (20%) 3. Terrorism (19%) 4. Iraq (15%) 5. Health care (8%) 6. Taxes (5%) 7. Education (4%) Los Angeles Times exit poll b 1. Moral/ethical values (40%) 2. Economy/jobs (33%) 3. Terrorism (45%) 4. Iraq (%) 5. Education (15%) 5. Social issues c (15%) 7. Health care (9%) 7. Taxes (9%) 9. Foreign affairs (5%) 9. Social Security (5%) 11. Medicare/Rx drugs (3%) 1. Moral values (35%) 2. Terrorism (32%) 3. Iraq (8%) 4. Economy/jobs (7%) 5. Taxes (6%) 6. Health care (4%) 7. Education (2%) 1. Moral/ethical values (54%) 2. Terrorism (45%) 3. Economy/jobs (%) 4. Social issues c (14%) 5. Education (12%) 6. Taxes (11%) 6. Iraq (11%) 8. Health care (5%) 9. Foreign affairs (3%) 9. Social Security (3%) 11. Medicare/Rx drugs (2%) 1. Economy/jobs (33%) 2. Iraq (22%) 3. Health care (13%) 4. Moral values (8%) 5. Education (7%) 6. Terrorism (5%) 6. Taxes (5%) 1. Economy/jobs (48%) 2. Moral/ethical values (24%) 3. Iraq (21%) 4. Education (%) 5. Social issues c (15%) 6. Health care (14%) 7. Terrorism (13%) 8. Foreign affairs (8%) 9. Taxes (7%) 9. Social Security (7%) 11. Medicare/Rx drugs (4%) SOURCES: National Election Pool Election Day exit poll (2 November 2004); and Los Angeles Times national Election Day exit poll (2 November 2004). a In the National Election Pool exit poll, voters could choose only one of seven items. Some voters did not choose any. b In the Los Angeles Times exit poll, voters could choose two of eleven items. c The survey offered the option of Social issues such as abortion and gay marriage. W March 2005

4 Health & Election for Kerry. However, the uninsured made up only 8 percent of those who said that they had voted, slightly more than half their proportion in the overall adult population. 14 Health care as a concern. Despite the fact that voters did not report health care as a major issue influencing their vote, health care remained an important concern among them. In the National Election Pool exit poll, 70 percent said that they were very concerned about the availability and cost of health care, and another 23 percent said that they were somewhat concerned. This concern was prominent for both Kerry voters (83 percent) and Bush voters (57 percent). 15 In addition, just a few weeks before the election, 47 percent of registered voters said that they were very worried about having to pay more for health care and insurance, outranking worries about not being able to pay rent or mortgage (25 percent), being the victim of a terrorist attack (21 percent), and losing a job (17 percent of those who were employed). Which health care issue was most important? When asked two weeks beforetheelectiontonameintheirownwordsthemostimportanthealthcareissuein deciding their vote for president, voters cited health care costs the most frequently, named by 30 percent of registered voters. The next most important health care issues were the uninsured and access to health care ( percent) and Medicare and seniors health care ( percent). Of note, Republican registered voters were more likely than Democrats (33 percent to 19 percent) not to name any health care issue as most important in the voting decision. 17 Registered voters were given a list of specific health and health care issues and asked how important they were in their vote for president. The list included some issues, such as abortion and gun control, that some think of as public health issues and others see as social issues. The top four health care issues for Republican registered voters were abortion (30 percent extremely important), bioterrorism (28 percent), health care and insurance costs (20 percent), and medical malpractice (20 percent). The top four issues for Democratic registered voters were the uninsured (43 percent), health care and insurance costs (39 percent), prescription drug benefits for seniors (39 percent), and the cost of prescription drugs (38 percent). Republican registered voters were less likely than Democrats to say that ten of the thirteen issues asked about were extremely important in deciding their vote (Exhibit 2). As might be expected, Independents generally fell between Republican and Democratic registered voters when rating the importance of these issues. 19 Thetopfourhealthcareissuesforvotersoverallremainedthesamethroughout the summer and fall campaign. 20 Apostelectionsurveyofself-describedBushvotersshowedthatamongthese voters, abortion ranked higher than other social issues as a voting issue. Nearly one-thirdsaidthatabortion(20percent)or partial-birth abortion(11percent) was one of the two most important issues in their voting decision, placing the issue ahead of gay marriage ( percent), stem cell research (5 percent), and school HEALTH AFFAIRS ~ Web Exclusive W5-89

5 EXHIBIT 2 Importance Of Health Care Issues In Deciding Preferences In The 2004 Presidential Election, By Party Identification, 2004 Percent saying extremely important Health care issue Cost of health care and health insurance Number of Americans without health insurance Cost of prescription drugs Prescription drug benefits for seniors Bioterrorism Medicare Problems with health care quality Abortion Gun control Racial disparities in health care Medical malpractice Stem cell research Aid to developing countries to prevent and treat HIV/AIDS prayer (3 percent). 21 Total registered voters Republicans Democrats Trusttohandlehealthcareissues.Polls indicate that if 2004 had been an election mainly about health care, Senator Kerry might have had an advantage. In the days before the 2004 election, likely voters were more likely to trust Kerry than Bush when it came to handling health care issues, by an average of 50 percent to 40 percent. 22 On the other hand, this is a smaller lead than Democrats have usually held. In February 2004, registered voters gave Democrats a twenty-one-point lead over Republicans (52 percent to 31 percent) on which party would do a better job on the health care issue. 23 Registered voters as a whole said they trusted Kerry more than Bush to handle seven of thirteen specific health and health care issues. Kerry s biggest advantages were on the uninsured, racial disparities in health care, and prescription drug costs. Bush held a lead on two health issues: bioterrorism and gun control. 24 Neither candidate s health care proposals generated much enthusiasm or anger. Only 9 percent of registered voters were enthusiastic about President Bush s policies, and 14 percent were enthusiastic about Senator Kerry s health care proposals. Sixteen percent of registered voters were angry with the president s policies, and 9 percent were angry about Kerry s proposals. 25 Republican versus Democratic voters health care concerns. The differences in Republican and Democratic voters interest in health care issues in the 2004 election appear to reflect major differences in their concerns about their own and ** SOURCE: Henry J. Kaiser Family Foundation/Harvard School of Public Health poll (October 2004). **p < ** 43** 38** 39** 26 35** 30** 27** 27** 17 22** 19** W March 2005

6 Health & Election thecountry shealthcareproblems.atthetimeoftheelection,republicanswere half as likely as Democrats to be very worried about several health care problems in their own lives (Exhibit 3). Two-thirds of Democratic registered voters (63 percent), compared with 27 percent of Republicans, were very worried about having to pay more for their health care or insurance. Nearly half of Democratic voters were very worriedaboutnotbeingabletoaffordthehealthcareservices(46percent)orprescription drugs (45 percent) they might need, compared with about one in five Republican voters. Among registered voters who had health insurance, 39 percent of Democrats were very worried about losing their health insurance, compared with only percent of Republicans. 26 Looking at the country as a whole, Republican voters (13 percent) were much less likely than Democratic voters (53 percent) to say they were not at all satisfied with the availability and affordability of health care in this country. While only percent of Republican voters were dissatisfied with the quality of health care in this country, 62 percent of Democrats said that they were dissatisfied. 27 In addition to the traditional health care issues that have arisen in past presidential elections, a new one to emerge is bioterrorism. Far more Democratic (58 percent) than Republican (9 percent) registered voters were dissatisfied with the country s preparedness to deal with bioterrorism. 28 Although they may have many policy differences, Republican and Democratic registered voters saw the same reasons behind rising health care costs. In November 2004, registered voters ranked high profits by drug and insurance companies EXHIBIT 3 Health Care Worries Among Registered Voters, By Party Identification, 2004 Downloaded from by Health Affairs Health care worry Having to pay more for your health care or insurance Not being able to afford health care services you think you need Not being able to afford prescription drugs you need Your health plan being more concerned about saving money than what treatment is best for you (among insured) Quality of health care services you receive getting worse Losing your health insurance (among insured) Having to stay in your current job instead of taking a new job for fear of losing health benefits (among those employed and insured) Percent very worried Total registered voters Republicans Democrats 63** 46** 45** 44** 37** 39** 32** on October 6, 2017 by HW Team SOURCE: Henry J. Kaiser Family Foundation/Harvard School of Public Health poll (October 2004). **p <.05 HEALTH AFFAIRS ~ Web Exclusive W5-91

7 (30 percent) and the number of malpractice lawsuits (23 percent) as the most important among seven possible causes for those rising costs (Exhibit 4). 29 Partisan Differences On Specific Health Care Issues Republican and Democratic voters held different views of what to do about specific health care issues. The role of government. When one is looking at health care as an issue, it is important to keep in mind that Republicans and Democrats generally differ in their views about the role of government in solving problems. For example, in the National Election Pool exit poll, Republican voters expressed considerably less interest than Democratic voters in expanding the role of government in trying to solve problems. Two-thirds (66 percent) of Bush voters said that government is doing too many things that are better left to businesses and individuals, while only 29 percent thought that government should do more to solve problems. About two-thirds (65 percent) of Kerry voters thought that government should do more. 30 Medicare prescription drugs. Democratic registered voters were far more critical (52 percent) than Republicans (20 percent) of the new Medicare prescription drug bill. The numbers were essentially reversed when it came to favorable attitudes. Many older people reported concern about the legislation. Two weeks before the 2004 election, 44 percent of seniors had an unfavorable impression of the bill, while27percenthadafavorableview,and8percentwereneutral.oneinfivese- EXHIBIT 4 Registered Voters Views About The Most Important Causes Of Rising Health Care Costs, By Party Identification, 2004 Downloaded from by Health Affairs Cause of rising costs High profits made by drug and insurance companies Number of malpractice lawsuits Greed and waste in health care system Aging of population Use of expensive, high-tech medical equipment and expensive new drugs Doctors making too much money Most people with health insurance have little incentive to look for lower-price doctors and services Other/none of these/all equally important (volunteered answer) Don t know/refused Percent saying most important Total registered voters Republicans ** Democrats on October 6, 2017 by HW Team SOURCE: Henry J. Kaiser Family Foundation/Harvard School of Public Health poll (October 2004). **p <.05 W March 2005

8 Health & Election niors (21 percent) did not have an opinion. 31 However, even though many seniors were dissatisfied with the law, President Bush won a majority of the senior vote, which suggests that other issues were more important to older voters. 32 The uninsured. The problem of the uninsured was less important as a voting issue for Republicans (17 percent of registered voters said that it would be extremely important in deciding their vote) than for Democrats (43 percent). 33 Prior polls showed that people in both parties think that something should be done about the problem, but Republicans (37 percent) were much less likely than Democrats (61 percent) to support a large-scale effort that might require a tax increase. 34 Abortion. In a survey just before the 2004 presidential election, abortion was the top health care issue for Republican voters. 35 A majority of Bush voters (62 percent) believed that abortion should be always or mostly illegal, while a majority of Kerry voters (74 percent) believed that abortion should be always or mostly legal. Nearly three-fourths (74 percent) of those who thought abortion should be always or mostly illegal voted for President Bush. 36 Drug importation. Onenationalhealthcareissuethatreceivedconsiderable attention during the campaign that did not show the same divide along partisan lines was drug importation. More than eight in ten (81 percent) registered voters favored changing the law to allow Americans to buy prescription drugs from pharmacies in Canada to get a lower price. A majority of both Republican (65 percent) and Democratic (90 percent) registered voters favored drug importation. 37 Health policy propositions in California. Two important health policy issues were voted on as ballot propositions in California, a state with 10 percent of the country s voters. In both cases, the partisan divide was evident. Proposition 72, a plan to deal with the uninsured problem by requiring most businesses to offer health insurance to their workers, lost in a close vote, 51 percent to 49 percent. Most Californians who voted for Bush in the presidential race (76 percent) voted against Proposition 72, while most who voted for Kerry (70 percent) voted for it. 38 Proposition 71, to enlarge the scope of future stem cell research and have the state put up $3 billion to support it, passed 59 percent to 41 percent. A majority of Bush supporters (63 percent) voted against the stem cell plan, while a majority of Kerry backers (78 percent) voted for it. 39 Discussion At the time of the 2004 presidential election, many voters were dissatisfied with aspects of the U.S. health care system and worried about health care related problems in their own lives. But the candidates positions on health care issues had relatively little effect on voting behavior, either because the issues were not seen as salient enough or because the candidates positions were not articulated well enough. In the end, most voters did not make their presidential voting decision based on health care issues. With concerns about future presidential leadership and issues of moral values, the economy, terrorism, and Iraq on their HEALTH AFFAIRS ~ Web Exclusive W5-93

9 With other concerns on their minds, most voters postponed their concerns about health care in considering their vote. minds, most voters postponed their concerns about health care in considering their vote. President Bush has named Social Security and taxes as his top second-term priorities, and malpractice reform as his number-one health care priority for the next year. This reflects at least in part the fact that Republican voters care more about this issue than Democratic voters do, as well as the fact that the president has had a long-term interest in malpractice reform going back to his days as governor of Texas. 40 In addition, the large federal deficit and the president s commitment to making his first-term tax cuts permanent means that it is extremely unlikely that substantial new money will be available for costly health initiatives such as covering the uninsured. For these reasons, no sweeping health reform proposal or large-scale effort to expand coverage for the uninsured is likely to gain serious consideration in the administration and Congress in the next few years. Many do believe, however, that the search for savings to reduce the federal deficit could result in new efforts to limit federal spending for both Medicare and Medicaid, and possibly in proposals to fundamentally restructure the Medicaid program, which now provides health insurance to some fifty million low-income, elderly, and disabled Americans. The Medicaid debate will likely unfold as part of the budget process, but, driven by the budget process and not by public opinion, it is likely to be the most consequential healthcareissueaddressedoncapitolhillinthenextyearintermsofthenumber of people affected and the dollars involved. The question for the future is whether health care will once again make the transition from being a problem about which people have deep concerns to a voting issue, as it did in the 1992 presidential election. 41 Thismaydependinparton whether health care can be framed as an issue of fundamental principles, rather than a series of complex health financing plans. That it is not now a voting issue suggests that politicians will not feel compelled to mount major health care reform efforts, at least as long as other issues dominate. But the fact that health care costs and access are such strong personal and family worries also suggests that candidates and politicians who decide to champion health care issues can win favor with the American people for doing so. The fact that personal concern about health care costs and access is already so high also suggests that raising awareness about these issues is not the paramount challenge for supporters of health care reform. Rather, the biggest challenge is to forge consensus on policy solutions and ways to pay for them that politicians can effectively champion. W March 2005

10 Health & Election Robert Blendon is a member of the Board of Directors of Assurant, a multi-product insurance company. NOTES 1. K. Victor and M.W. Serafini, The Senate A Tighter Grip, National Journal (6 November 2004): M.L. Amer, Membership of the 109th Congress: A Profile, Congressional Research Service, 20 December 2004, (17 February 2005). 3. P.H. Conley, Presidential Mandates: How Elections Shape the National Agenda (Chicago: University of Chicago Press, 2001); W.D. Burnham, Bill Clinton: Riding the Tiger, in G.M. Pomper et al., TheElectionof1996:Reports and Interpretations (Chatham, N.J.: Chatham House, 1997), 10 11; and J.H. Aldritch Jr., Why Parties? The Origin and Transformation of Political Parties in America (Chicago: University of Chicago Press, 1995), Henry J. Kaiser Family Foundation/Harvard School of Public Health poll, October 2004 (Storrs, Conn.: Roper Center for Public Opinion Research, 2004). 5. This methodology for telephone surveys is described in R.J. Blendon et al., Americans Health Priorities Revisited after September 11, 13 November 2001, content.healthaffairs.org/cgi/content/abstract/hlthaff.w1.96 (20 January 2005). The margin of error for a survey of 800 respondents is plus or minus four percentage points; for 1,200 respondents, plus or minus three percentage points. 6. Los Angeles Times National Election Day exit poll, 2 November 2004 (Storrs, Conn.: Roper Center, 2004); Los Angeles Times California Election Day exit poll, 2 November 2004 (Storrs, Conn.: Roper Center, 2004); and National Election Pool Election Day exit poll, 2 November 2004 (Storrs, Conn.: Roper Center, 2004). The margin of error for each of these exit polls is plus or minus three percentage points. 7. Gallup/CNN/USA Today poll, October 2004 (Storrs, Conn.: Roper Center, 2004); and Kaiser/Harvard poll, 4 28 November 2004 (Storrs, Conn.: Roper Center, 2004). 8. Pew Research Center/Princeton Survey Research Associates poll, 5 8 November 2004, people-press.org/reports/display.php3?reportid=233 (20 January 2005); and Harris poll, 8 15 December 2004 (Rochester, N.Y.: Harris Interactive, 2004). 9. National Election Pool Election Day exit poll (2 November 2004); and Los Angeles Times National Election Day exit poll (2 November 2004). 10. Newsweek/Princeton Survey Research Associates poll, January 2004 (Storrs, Conn.: Roper Center, 2004). 11. National Election Pool Election Day exit poll (2 November 2004); and Los Angeles Times National Election Day Exit Poll (2 November 2004). 12. National Election Pool Election Day exit poll (2 November 2004). 13. Ibid. 14. Kaiser/Harvard poll (4 28 November 2004). 15. National Election Pool Election Day exit poll (2 November 2004).. Kaiser/Harvard poll (14 17 October 2004). 17. Ibid.. Ibid. 19. Ibid. 20. R.J. Blendon et al., Health Care in the 2004 Presidential Election, New England Journal of Medicine 351, no. 13 (2004): Fabrizio McLaughlin poll, November 2004 (Storrs, Conn.: Roper Center, 2004). 22. Fox News/Opinion Dynamics poll, October 2004 (Storrs, Conn.: Roper Center, 2004); Newsweek/ PSRA poll (27 29 October 2004); and Los Angeles Times poll, October 2004 (Storrs, Conn.: Roper Center, 2004). 23. Fox News/Opinion Dynamics poll, 19 February 2004 (Storrs, Conn.: Roper Center, 2004). 24. Kaiser/Harvard poll (14 17 October 2004). 25. Ibid. 26. Ibid. 27. Ibid. HEALTH AFFAIRS ~ Web Exclusive W5-95

11 28. Ibid. 29. Kaiser/Harvard poll (4 28 November 2004). 30. National Election Pool Election Day exit poll (2 November 2004). 31. Kaiser/Harvard poll (14 17 October 2004). 32. National Election Pool Election Day exit poll (2 November 2004); and Los Angeles Times National Election Day exit poll (2 November 2004). 33. Kaiser/Harvard poll (14 17 October 2004). 34. Harvard School of Public Health/Robert Wood Johnson Foundation/International Communications Research poll, 30 May 3 June 2003 (Storrs, Conn.: Roper Center, 2003). 35. Kaiser/Harvard poll (14 17 October 2004). 36. National Election Pool Election Day exit poll (2 November 2004). 37. Kaiser/Harvard poll (14 17 October 2004). 38. Los Angeles Times California Election Day exit poll (2 November 2004). 39. Ibid. 40. N. Lemann, The Controller, in The Best American Political Writing 2004, ed. R. Flippin(New York: Thinder s Mouth Press, 2004). 41. R.J. Blendon et al., The Implications of the 1992 Presidential Election for Health Care Reform, Journal of the American Medical Association 268, no. 23 (1992): Downloaded from by Health Affairs on October 6, 2017 by HW Team W March 2005