MEDICAL MALPRACTICE

Spending

Certain changes in medical liability law could reduce spending:

• Caps on non–economic damages reduce the average payout per medical malpractice claim.
• There is limited evidence about how other changes in medical liability law affect the number and average payout of claims or MM insurance premiums.
• Changing medical malpractice law probably will have only a small direct effect on overall health care spending.

Caps on non-economic damages reduce the average payout per medical malpractice claim.

Changes in medical liability law are intended to reduce the payout and/or number of claims. In principle, these reductions should translate into lower total payout in MM damages and settlements, ultimately lowering MM insurance costs and related premiums and reducing incentives for physicians to practice “defensive medicine.”

Overall, the literature suggests that non-economic damage caps are associated with lower average payouts per MM claim. The literature is equivocal about a link between caps on non–economic damages and number of claims.

The following studies are differentiated as being “strong” or “weak.” Mello (2006) explains the distinction as involving a series of methodological considerations, including use of appropriate data sources, low potential for bias, use of appropriate analytical methods, adequate control for confounding variables, and adequacy of sample size. We applied similar criteria in reviewing the strength of more recent empirical studies from 2005 to 2007. Mello does not distinguish between studies that compared different kinds of caps (e.g., non–economic damage caps versus total damage caps).

Effects of non–economic damage caps on average payout per claim. Mello’s review of the empirical literature through 2005 identified three strong studies (Danzon, 1984; Danzon, 1986; Sloan, Mergenhagen, and Bovbjerg, 1989) and two weak studies suggesting that caps were associated with lower average payout per claim. One strong study (Zuckerman, Bovbjerg, and Sloan, 1990) and one weak study failed to show an effect (Mello, 2006).

Findings from the more recent literature (2005—2007) are basically consistent with earlier assessments. Three strong studies (Avraham, 2007; Guirguis–Blake et al., 2006; Waters et al., 2007) and one weak study (Frech, Hamm, and Wazzan, 2006) found that caps on non–economic damages were associated with lower average payout per claim. More specifically, Guirguis–Blake et al. reported that average payout per claim was 22 percent less in states with non–economic damage caps than in states without the caps; Waters et al. indicated that non–economic damage caps were associated with a reduction in average payout per claim of almost $15,000; and Avraham found that non–economic damage caps reduced average payouts by between 65 and 74 percent across several different regression models.

Effects of non-economic damage caps on number of claims. Mello identified only one strong study that examined how caps affect the number of MM claims; the study found no effect (Zuckerman, Bovbjerg, and Sloan, 1990). Three more–recent studies (Avraham, 2007; Frech, Hamm, and Wazzan, 2006; Waters et al., 2007) reported that the caps were associated with fewer claims. However, another (strong) study found no evidence of a link between non–economic damage caps and number of claims (Donohue and Ho, 2007).

Effects of non–economic damage caps on MM insurance premiums. The extent to which any decreases in MM liability costs would be translated into lower MM insurance premiums may depend on the type of malpractice insurer. We would expect that cost savings of insurance companies that are owned by physicians would be passed on to their insureds through lower premiums or payments of dividends. In contrast, the extent to which investor owned insurers pass on cost reductions may depend on the degree of competition among insurers in a state.

Studies of how caps on non–economic damages affect MM insurance premiums have mixed results. Mello’s review of the empirical literature through 2005 identified four studies (two of them strong: Zuckerman, Bovbjerg, and Sloan, 1990; Thorpe, 2004) showing a link between caps and lower insurance premiums, and four more studies (one of them strong: Zuckerman, Bovbjerg, and Sloan, 1990) failing to show a link.

Effects of non–economic damage caps on aggregate health care expenditures. One recent study examined the direct relationship between MM damage caps and a measure of aggregate health care expenditures. Hellinger and Encinosa (2006) looked at the association between state non–economic damage caps and per capita state level health care expenditures between 1984 and 1998. They found that the presence of a cap was significantly associated with states’ having lower average per capita health care expenditures (by about 3 or 4 percent) during a specified period. However, we were unable to classify this as a methodologically strong study.

There is limited evidence about how other changes in medical liability law affect the number and average payout of claims or MM insurance premiums.

Effects of limiting attorney fees. Mello’s review of the 1984—2005 literature identified five studies (four strong: Danzon, 1984; Danzon, 1986; Sloan, Mergenhagen, and Bovbjerg, 1989; Zuckerman, Bovbjerg, and Sloan, 1990) that looked for an association between limiting attorney fees and MM claims payout; one study (strong: Zuckerman, Bovbjerg, and Sloan, 1990) that looked for an association between limiting attorney fees and MM claims frequency; and four studies (two strong: Zuckerman, Bovbjerg, and Sloan, 1990; Thorpe, 2004) that looked for an association between limiting attorney fees and MM insurance premiums. None of these studies found a link between limiting attorney fees and these outcomes.

Two newer studies examined the relationship between limiting attorney fees and number or average payout of claims; neither study found evidence of a link (Guirguis–Blake et al., 2006; Waters et al., 2007). None of the newer studies examined how limiting attorney fees affected insurance premiums or losses.

Effects of modifying the collateral source rule. Mello’s review of the literature through 2005 identified five studies that looked for an association between modifying the collateral source rule and MM claims payments. Of those, two studies (both strong: Danzon, 1984; Danzon, 1986) found a negative effect; three studies (two of them strong: Sloan, Mergenhagen, and Bovbjerg, 1989; Zuckerman, Bovbjerg, and Sloan, 1990) found no effect.

Two other studies (both strong: Danzon, 1986; Zuckerman, Bovbjerg, and Sloan, 1990) looked for a link between modifying the collateral source rule and the number of MM claims. One study found that modifying the rule reduced the number of claims; the other study found no effect.

Finally, Mello identified five studies (two of them strong: Zuckerman, Bovbjerg, and Sloan, 1990; Thorpe, 2004) that looked for a connection between modifying the collateral source rule and lower MM insurance premiums. None of the studies found evidence to support an association.

Three newer studies (all of them strong) examined the association between modifying the collateral source rule and the number and average payout of MM claims. None of them found a relationship between modifying the rule and claims frequency; only one study found that rule modification was associated with lower claims payments (Avraham, 2007; Guirguis-Blake et al., 2006; Waters et al., 2007).

Two other studies (both strong) report that modifying the collateral source rule was only sometimes associated with lower MM premiums and long–run losses, for example, in connection with some medical specialties or insurance firms (Born, Viscusi, and Baker, 2006; Kilgore, Morrisey, and Nelson, 2006).

Effects of modifying joint and several liability. Mello’s literature review through 2005 identified two studies (both weak) that found no link between modifying the joint and several liability rule and MM claims. Mello found no studies that examined whether rule modification was associated with fewer MM claims. However, she identified four studies (one of them strong: Thorpe, 2004) that had looked for an association between modification of the joint and several liability rule and MM insurance premiums. The three weak studies found evidence that modifying the rule reduced premiums; the strong study found no link.

More recent studies yielded similarly equivocal results. Only one of three newer studies (all strong) showed that modifying the joint and several liability rule was associated with fewer MM claims; none showed an association with reduced payouts (Avraham, 2007; Guirguis–Blake et al., 2006; Waters et al., 2006). Two other newer studies (both strong) found that modifying the joint and several liability rule was not associated with lower MM premiums, but with lower, long–run MM insurance losses for some (but not all) insurance firms (Born, Viscusi, and Baker, 2006; Kilgore, Morrisey, and Nelson, 2006).

Effects of other rule changes. Other changes to the medical liability law have been tried (and even studied). They include caps on punitive damages and on total damages, periodic-payment interventions, modifications in statutes of limitations, and pretrial screening mechanisms. The evidence base concerning these changes ranges from limited to non-existent.

Changing medical malpractice law probably will have only a small direct effect on overall health care spending.

There is little evidence about the direct relationship between changes in liability law and broader measures of health care expenditures. The direct costs associated with MM claims and insurance likely represent less than 2 percent of total health care expenditures in the United States, thus bounding the potential for broader savings. Nevertheless, reductions in MM costs could plausibly contribute to savings in broader expenditures, especially if changes to liability law reduce the extent of defensive medicine.

Recent CBO estimates (CBO, 2009) suggest that a package of national tort interventions including caps on damages and joint-and-several liability reforms potentially could reduce total national spending on health care by 0.5% ($11B) in 2009. We are unable to identify or confirm the empirical basis for the latest CBO estimate. The most recent CBO conclusion notably represents a reversal from a previous CBO empirical study of malpractice tort reforms (CBO, 2006), which concluded that the impact of on health care spending of various types of statutory tort intervention has been highly varied, and that “the estimated [savings] effect of implementing a package of previously proposed tort limits is near zero.”

Several recent studies (2007 to 2009) have attempted to link malpractice pressure or tort reforms directly to aggregate health care expenditures. Avraham et al. (2009) found that several types of tort intervention, including caps on damages, were associated with reduced spending on employer-based health insurance premiums. In a similar study, however, Morrisey et al. (2008) found no such effect. Roberts & Hoch (2009) notably did find an association between a different measure of malpractice litigation pressure and Medicare Part B (outpatient) expenditures, with the former reportedly accounting for approximately 2.6% of the latter. Baicker et al. (2009) analyzed the relationship between malpractice insurance premiums and damage awards and Medicare expenditures for several specific categories of procedures, and concluded that increased malpractice pressure is associated with increased spending for some types of medical services (most notably, imaging services). Finally, Lakdawalla & Seabury (2009) found that only about 5% of observed growth in medical expenditures from 1985-2003 could be explained by variation in another measure of malpractice pressure (jury generosity), and that this effect on expenditures was likely balanced by reduced mortality rates, which were also found associated with increased malpractice pressure.

Consumer Financial Risk

There is no direct relationship between MM tort interventions and public financing for the health care sector, neither is there empirical evidence available to describe or quantify any such relationship.

Waste

Changing the medical liability law plausibly might reduce waste:

• Changing the medical liability law plausibly might reduce the waste associated with the practice of defensive medicine.
• Cost effects of defensive medicine are difficult to establish empirically.

Changing the medical liability law plausibly might reduce the waste associated with the practice of defensive medicine.

Medical malpractice (MM) liability may give physicians incentives to practice defensive medicine, such as ordering unnecessary medical tests or procedures primarily intended to avoid liability, rather than to benefit patients. Thus it is plausible that changing the liability law could reduce defensive medicine practices and, therefore, waste.

Some studies (e.g., Studdert et al., 2005) suggest that certain medical specialties (e.g., OB/GYN) and procedures (e.g., cesarean sections) are especially prone to medical liability pressure, making them reasonably reliable indicators or surrogates for defensive medicine more broadly.

In her review of empirical literature through 2005, Mello (2006) identified seven studies that looked for a relationship between medical liability pressure and rates of performance for medical procedures deemed vulnerable to defensive medicine (in particular, cesarean sections). According to Mello, three strong studies (Dubay, Kaestner, and Waidmann, 1999; Localio et al., 1993; Tussing and Wojtowycz, 1997) and one weak study found small but significant associations between higher cesarean section rates and increased medical liability pressure; three other weak studies found no association. Mello also identified three additional studies that examined other clinical indicators of defensive medicine. She concluded that research “consistently find[s] that assurance [defensive medicine] behaviors are widespread, and become even more so during malpractice crises.”

In a search of the 2005—2007 literature, only two newer studies investigating the effects of defensive medicine were found. Dhankhar, Khan, and Bagga (2007) looked at the relationship between medical liability pressure, health outcomes, and resource use in treatment for patients with acute myocardial infarction (AMI). They reported that increased medical liability pressure was actually associated with lower resource use and better clinical outcomes for at least some AMI patients. This result is seemingly at odds with widely cited work on defensive medicine by Kessler and McClellan (1996, 2002), which found the opposite effect. On a different note, Murthy et al. (2007), examining an Illinois data set from 1998 to 2003, found that rising county level cesarean section rates were associated with higher MM insurance premiums for gynecologists, a result that suggests a defensive medicine effect.

Cost effects of defensive medicine are difficult to establish empirically.

Another body of literature has attempted to quantify the costs associated with defensive medicine practices, without specific reference to medical malpractice. Fully reviewing the literature on cost estimates of such practices goes beyond the scope of this summary, but Baicker, Fisher, and Chandra (2007) offered a good recent example of a relevant empirical study. That study suggested associations between higher MM costs and insurance premiums, patterns in Medicare service usage, and, ultimately, increased Medicare spending. However, establishing the costs of defensive medicine in a robust and comprehensive way has proven very difficult (Mello, 2006).

The newest studies (2007 to 2009) to investigate the defensive medicine effects of malpractice pressure have generated mixed findings. Several recent studies have sought to identify links between malpractice pressure and broader healthcare expenditures and utilization (e.g., Avraham et al., 2009; Morrisey et al., 2008; Baicker et al., 2009; Robers & Hoch, 2009: Lakdawalla & Seabury, 2009). Findings from this research vary somewhat across studies, measures, and empirical designs. Several of these studies find evidence for significant defensive medicine effects, but those effects manifest in different ways in different studies, and in some cases appear to be limited to specific categories of medical procedures. At least one study touching on this issue failed to find a defensive medicine type of effect (Morrisey et al., 2008), and another of the recent studies to examine this issue (Lakdawalla & Seabury, 2009) suggested that an observed defensive medicine effect was counterbalanced by improved mortality rates for patients, which was also associated with increased malpractice pressure. Finally, several recent studies that focused on the relationship between malpractice pressure and obstetrical services actually found that increased malpractice pressure was associated both with increased performance of C-sections, and with improved outcomes for patients, in at least some circumstances.

Reliability

Little empirical evidence has been derived from studies that directly examined the association between MM tort interventions and conventional quality of care measures such as reliability.

Patient Experience

We know of no studies suggesting that MM tort interventions directly affect the quality of patient experience in the health care system.

Health

There is limited, and mixed, evidence about how changes in medical liability law affect patient health:

• Evidence about the effects of medical liability law on patient health is scant.
• Available evidence about the health effects of liability law is inconsistent.

Evidence about the effects of medical malpractice liability law on patient health is scant.

In principle, medical malpractice (MM) liability deters health care providers from negligent behavior that could cause injury. Thus, reducing MM liability might also reduce the incentive for health care providers to deliver appropriate medical treatment skillfully. Recent studies have shown that better care translates into better health outcomes, so reducing liability could degrade patient health.

However, reducing MM liability might also make physicians more willing to provide medically appropriate, but high risk, treatment. Thus, reducing liability could either improve or degrade the health of at least some patients; empirical evidence regarding either effect is very limited.

The deterrence effects of MM liability are very difficult to study rigorously (Mello and Brennan, 2002), and few studies have directly examined the association between medical liability pressure and health outcomes. In perhaps the most influential work in this area, Kessler and McClellan (1996, 2002) estimated the relationship between medical liability pressure and health by examining measurable health outcomes (i.e., one year mortality and hospital readmission rates) for elderly (Medicare) heart attack patients. They found no statistically significant association between medical liability pressure and these health outcomes. Several other researchers have examined a potential link between medical liability pressure and obstetrical health outcomes, and they also failed to find an association (see, e.g., Dubay, Kaestner, and Waidmann, 1999; Sloan et al., 1995).

Available evidence about the health effects of liability law is inconsistent.

Most recently, Konety et al. (2005) explored the effects of non-economic damage caps on treatment and outcomes for bladder cancer. They found that several regions in the United States with non–economic damage caps had higher rates of radical cystectomy procedures and better bladder cancer survival rates than did several comparison regions without non–economic damage caps. By contrast, Dhankhar, Khan, and Bagga (2007) studied the effects of medical liability pressure on health outcomes for heart attack patients. They found that increased medical liability pressure was associated with better health outcomes for patients with less severe cases. However, medical liability pressure was unrelated to health outcomes for more severe heart attack cases.

The newest studies (2007 to 2009) to investigate the deterrence effects of malpractice pressure on health outcomes have generated mixed findings. Lakdawalla and Seabury (2009) notably found a significant association between patient mortality rates and a measure of malpractice pressure (jury generosity), while Baicker et al. (2008) and Sloan and Shadle (2009) notably failed to identify similar broad effects on mortality, using different measures of malpractice pressure. Dhankhar & Khan (2007), Iizuka (2008) and Currie & MacLeoad (2008) all studied the impact of malpractice pressure on obstetrical procedures (C-sections) and outcomes, and while their findings are not fully consistent, all found some evidence that greater malpractice pressure is at least sometimes associated with superior obstetrical outcomes – a result that was interpreted as a deterrence effect of liability.

Coverage

There is limited evidence about how changes in medical liability law might affect health insurance coverage.

Sun and Schmit (2007) used data from 1997 to 2003 to look at the association between several types of MM interventions and the fraction of state residents who were without health insurance. They found that non–economic damage caps were significantly associated with a smaller fraction of uninsured residents. Counter to their expectations, they did not find any association between the fraction of uninsured and either modification of the collateral source rule or punitive damage caps. Similarly, Avraham and Schanzenbach (2007) drew on data from 1981 to 2004 to look at the association between several types of MM interventions and private health insurance coverage. They concluded that total damage caps were associated with higher rates of private health insurance coverage, at least among persons who are young, single, or self–employed. Although the methods in these studies are only moderately strong, the results are nevertheless interesting.

Capacity

There is limited—and mixed—evidence about how changes in medical liability law affect the capacity of the health care system:

• Strong studies have found little association between medical liability pressure and overall supply of physician services. Read more below
• Some studies have suggested an association between medical liability pressure and physician supply for some medical specialties, particularly in regions affected by “crisis growth” in malpractice premiums. Read more below

Strong studies have found little association between medical liability pressure and overall supply of physician services.

Heightened MM liability risk and costs have sometimes been associated with professional dissatisfaction among physicians and their (self–reported) reduced willingness to perform some kinds of specialized medical procedures. In principle, one would expect that very high levels of liability risk (and associated costs) might lead some physicians to stop practicing medicine. By extension, changes in medical liability laws that reduce liability pressures might make the labor market more attractive to physicians, thereby increasing the supply of physician services (and, therefore, capacity).

In practice, the empirical literature is limited, and findings have been mixed concerning the associations between MM liability pressure and physician supply. In her review of the empirical literature, Mello (2006) identified five studies that looked for an association between changes in medical liability law and physician supply. Four of those studies focused on damage caps: one strong study (see the definition of strong in the Spending section above) found no general association between caps and supply (Matsa, 2005); however, the study did find a link between the existence of caps and physician supply for surgeons and “support specialists” (anesthesiologists, neurologists, pathologists, psychiatrists, and radiologists) in rural areas. The three other (weaker) studies found that damage caps were associated with increased physician supply. However, two of these studies also generated peculiar or counterintuitive results (e.g., higher–dollar damage caps were associated with more physician supply than were more restrictive caps). One other strong study (Kessler, Sage, and Becker, 2005) found that increased physician supply was associated with the existence of non–economic damage caps, but the association was not consistent across medical specialties.

Mello (2006) also reviewed three studies that looked for correlations between other indicators of medical liability pressure (e.g., levels of MM insurance premiums and claims payout) and physician supply. Two of the studies (one strong: Baicker and Chandra, 2005) found no association between medical liability pressure and physician supply; the third (weak) study found that higher MM insurance premiums were associated with reduced physician supply.

In more recent literature, only three studies focused on the relationship between physician supply and medical liability pressure or changes in medical liability law. Helland and Showalter (2006) found that a 10 percent increase in malpractice premiums was associated with nearly a 3 percent reduction in hours worked by physicians.

Some studies have suggested an association between medical liability pressure and physician supply for some medical specialties, particularly in regions affected by “crisis growth” in malpractice premiums.

For example, Mello et al. (2007) examined effects on physician supply in Pennsylvania during a crisis period of rapid growth in malpractice insurance premiums from 1999 to 2001 (compared with a baseline from 1993 to 1998). They found little evidence to support a broad association between medical liability pressure and specialist physician supply, or between medical liability pressure and specialists’ avoidance of high risk clinical procedures. However, Mello et al. did find evidence that the supply of OB/GYNs in Pennsylvania shrank by 8 percent during the period identified as a malpractice crisis.

Most recently, Yang et al. (2008) investigated the national supply of OB/GYNs as a function of medical liability pressure throughout the United States during 1992—2002. They found no broad association in their data between OB/GYN supply and several indicators of medical liability pressure.

Operational Feasibility

There is no broad legal impediment to changing medical liability law:

• Changes in medical liability rules have already been implemented by many state legislatures.
• Implementing statutory tort interventions does not require operational activity by health care organizations, but responding to the new laws potentially does.

Changes in medical liability rules have already been implemented by many state legislatures.

A number of states have adopted a range of statutory MM interventions, including caps on non–economic damages and limits on attorney contingency fees. Although particular interventions in specific states have sometimes been ruled unconstitutional by state courts, there is no general legal impediment against enacting laws to modify MM tort liability.

Implementing statutory tort interventions does not require operational activity by health care organizations, but responding to the new laws potentially does.

Enacting new statutes per se does not require insurers, health care organizations, or providers to do anything, although responding to new or changing legal requirements may have financial and operational costs. For example, costs could accrue for insurers who need to redo their actuarial models to accommodate how legal changes might affect risk management.