Does Sorry Work the Impact of Apology Laws on Medical Malpractice JRU final 3

30 Appendix Proof of Proposition I Part (a) and Part (b) The patient decides to settle if the benefit of settling, , is greater than the benefit of going to trial, . The probability of settling is given by: , , Pr The doctor proposes a settlement – (3) to minimize her expected payment: , 1– (4) Minimizing the doctor’s malpractice costs from Equation (2) using the probability of settlement given by Equation (3) yields the optimal settlement offer: , (5) 2 Doctors offer a higher settlement when their costs of going to court are higher and a lower settlement when the costs the patient faces are higher. The optimal settlement probability by the patient shows that the patient is more likely to settle as the costs of going to court rise: (6) 2 The patient’s probability of litigating, , is then given by the probability that the expected malpractice payment is greater than the psychic cost of litigating: Pr 1– , – (7) Using the assumption that psychic costs follow a uniform distribution and that we have an interior solution, we can reduce this probability to: , Again, . (8) is the additional psychic cost for the patient to sue if the physician apologizes. Consistent with the empirical evidence (Sloan & Hsieh, 1995), the probability of litigation given in Equation (8) is increasing with more serious health outcomes, decreasing in the costs of going to trial, but increasing in the probability an early settlement is reached. 31 Combining these results allows us to write the closed form solution for the expected malpractice payments net of costs for patient and doctor: , , Net Gain for Patient: Π (9) Net Cost for Doctor: Π , , Finally, consider the doctor’s incentives to apologize. The doctor will apologize for all health outcomes where ΠD ,1 Π , 0 :29 Pr ,1 : ΠD Π ,0 (10) From Equation (6) we can calculate the difference in settlement probabilities after an apology to see that settlements increase in the event of an apology: (11) 2 However, the effect of an apology on the likelihood of initiating litigation depends on the relative | | effect of the apology on the psychic costs which makes litigation less attractive, with the effect of the apology on settlement probabilities and judgment payments which makes litigation more attractive. From Equation (8) the effect of an apology on probability to litigate is given by: | | ,1 ,0 2 (12) 2 The effect of an apology on the probability to litigate is increasing in the effect on judgment sizes— ,1 , 0 —and decreasing in the psychic costs an apology imposes, . Perhaps more interestingly, apologies make patients more likely to litigate when the costs of going to court (both actual and psychic) are higher due to the fact that one deterrent to litigation is the threat of having to pay high court costs, and apologies reduce the likelihood of going to court in the event of litigation. 29 Note that once the health outcome is realized, doctors will apologize deterministically. , therefore, represents the ex ante probability that the doctor will apologize. We include this expression for ex ante probability of apology since it will be useful for discussing moral hazard and welfare in the next section. 32 Proof of Proposition 2 Part (a) We can see from Equation (6) that the apology law reduces the expected payment in case of an apology, Π , 1 , but has no effect on expected payments when no apology is made, Π health outcomes for which the doctor would apologize, :Π ,1 Π , 0 , so the set of , 0 , must be larger than before the laws were passed. Part (b) From Equation (8), a patient decides to initiate litigation if the expected benefit from litigation outweighs the costs of litigation. Apology laws reduce judgment sizes which decreases the benefits of litigation; and thus, the probability that the patient litigates decreases. Part (c) From the probability of settlement given in Equation (6), the likelihood of settlement is always higher in the event of an apology. Since apologies are more frequent, we expect more settlements. Part (d) It can be seen from Equation (5) that settlements are smaller in the event of an apology (which are now more common) and smaller still after a law reduces ,1 . Part (e) Since the laws increase settlement, reduce probability of litigation, reduce both judgment and settlement sizes, then we see from Equation (2) that malpractice payments net of costs made by the doctor must also go down. 33 Part (f) Given symmetric information and risk neutral parties, the welfare implication of the law is unambiguous: since we assume that doctor effort is unaffected, the only effect of litigation is a transfer from the defendant to the plaintiff that imposes a deadweight loss from the cost of litigation ( ). Thus the reduced likelihood of litigation and judgment means that the law increases welfare. 34 4000 3500 Number of Claims Resolved 3000 2500 2000 1500 1000 500 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Time to Resolution (years) Figure 1: Histogram of Claims By Time to Resolution 15 16 Table 1. State with Statutes Pertaining to Apology Law State Year Law Full Versus Passed Partial Statutes Massachusetts Texas California Florida Washington Tennessee Colorado Oregon Maryland North Carolina Ohio Oklahoma Wyoming Connecticut Louisiana Maine Missouri New Hampshire South Dakota Virginia Arizona Georgia Illinois Montana West Virginia Delaware Idaho Indiana Iowa South Carolina Utah Vermont Hawaii Nebraska North Dakota District of Columbia 1986 1999 2000 2001 2002 2003 2003 2003 2004 2004 2004 2004 2004 2005 2005 2005 2005 2005 2005 2005 2005 2005 2005 2005 2005 2006 2006 2006 2006 2006 2006 2006 2006 2007 2007 2007 ALM GL ch. 233, § 23D (1986) Tex Civ Prac & Rem Code Ann 18.061 (1999). Cal Evid Code 1160 (2000). Fla Stat Ann Ch 90.4026 (2004). Rev. Code Wash. §5.66.010(2002) Tenn. Evid. Rule §409.1 Colo Rev Stat Sec 13-25-135 (2003) Oreg Rev Stat Sec 677.082 (2003). Md. COURTS AND JUDICIAL PROCEEDINGS Code Ann. § 10-920 N.C. Gen. Stat. § 8C-1, Rule 413 (2004) ORC Ann. 2317.43 (2006) (63 Okl. St. § 1-1708.1H Wyo Stat. § 1-1-130 Conn. Gen. Stat. § 52-184d (2005) La. R.S. 13:3715.5 (2005) 24 M.R.S. § 2907 (2005) Mo.Rev.Stat §538.229 (2005) N.H.Rev. Stat. Ann. § 507-E:4 (2005) S.D. Codified Laws § 19-12-14 (2005) Va. Code Ann. §8.01-581.20:1 (2005) A.R.S. § 12-2605 O.C.G.A. § 24-3-37.1 735 ILCS 5/8-1901 (2005) Mont. Code Anno., § 26-1-814 (2005) W. Va. Code § 55-7-11a (2005) Delaware Del. Code Ann. Tit. 10, 4318 (2006) Ida. ALS 204; 2006 Idaho Sess. Laws 204; Ind. HEA 1112 Iowa HF 2716 (2006) South Carolina Ch.1, Title19 Code of Laws 1976, 19-1-190 (2006) 2006 Ut. SB 41 Vermont S 198 Sec. 1. 12 V.S.A. 1912 (2006) HRS section 626-1, Hawaii Rules of Evidence Rule 409.5 Nebraska Neb. Laws L.B. 373 (2007) North Dakota ND H.B. 1333 (2007) D.C. Code 16-2841 (2007) Partial Partial Partial Partial Partial Partial Full Partial Partial Partial Partial Partial Partial Full Partial Partial Partial Partial Partial Partial Full Full Partial Partial Partial Partial Partial Partial Partial Full Partial Partial Partial Partial Partial Partial Table 2. Variable definition and NPDB Summary Statistics Variable name Variable Definition Panel A: Individual Level Claim Payout Average Claim Payout in Y2000 dollars Years to Resolution Average Year to Reach Resolution Outcomes (Available If Reported After 2004) Emotional Injury A dummy variable which takes the value 1 if the medical outcome is emotion injury and 0 otherwise Insignificant Injury Mean $200120 3.86 0.02 A dummy variable which takes the value 1 if the medical outcome is insignificant injury and 0 otherwise 0.03 Minor Temporary Injury A dummy variable which takes the value 1 if the medical outcome is minor temporary injury and 0 otherwise 0.15 Major Temporary Injury A dummy variable which takes the value 1 if the medical outcome is major temporary injury and 0 otherwise 0.09 Minor Permanent Injury Sign. Permanent Injury Insig. injury Somewhat A dummy variable which takes the value 1 if the medical Sign. outcome is minor permanent injury and 0 otherwise injury A dummy variable which takes the value 1 if the medical outcome is sig. permanent injury and 0 otherwise Major Permanent Injury 0.14 0.14 A dummy variable which takes the value 1 if the medical outcome is major permanent injury and 0 otherwise 0.09 A dummy variable which takes the value 1 if the medical outcome is quadriplegic and 0 otherwise 0.04 Death A dummy variable which takes the value 1 if the medical outcome is death and 0 otherwise 0.28 Cannot be Determined A dummy variable which takes the value 1 if the medical outcome cannot be determined and 0 otherwise 0.02 A dummy variable which takes the value 1 if the payment is a result of settlement and 0 otherwise 0.90 Judgment A dummy variable which takes the value 1 if the payment is a result of judgment and 0 otherwise 0.03 Unknown A dummy variable which takes the value 1 if the payment type is unknown and 0 otherwise 0.07 Quadriplegic Paymentt T P Type Settlement Panel B: State Level Claims Frequencies Claim Payments Sig. Injury Average state-level claim frequencies for incidents that took place in 2000 Average state-level claim payment in 2000 (millions of Y2000 dollars) 316.86 71.33 Table 3. The Impact of Apology Law on Medical Malpractice Claims Frequencies and Claims Severity Dependent Variable log (Claim Frequencies) (1) Adopted Apology Law (2) 0.142 (0.086) Other Law Changea (3) 0.153 (0.083)* (1) 0.147 (0.095) X X b log (Claim Payout) (2) 0.279 (0.163)* (3) 0.276 (0.163)* 0.202 (0.181) X X Other Covariates State Fixed Effects Year Fixed Effects X X X X X X X X X X X X X X Observation R-squared 867 0.97 867 0.97 867 0.97 867 0.97 867 0.97 867 0.97 Note: * significant at 10%; ** significant at 5%; *** significant at 1%. Unit of Observation is StateYear. Each column shows the results from a separate diff-in-diff regressions. Standard errors are clustered at the state level. The dependent variables are either Log (Number of Cases) in a state-year or Log (Total amount of Settlement) in a state-year. a. Other law change includes non-economic damage cap, punitive damage cap, law on medical malpractice disclosure, csr_tort and jsl_tort. b. Covariates include population, % Black, % White, % of population that are 65 or above, and # of Physicians. Adopted Apology Law Before 2002 Table 4. Long Term Impact of Apology Laws on Claim Frequencies in Early Adoption States Dependent Variable: Log (Malpractice Claim Frequencies in State-Year) (1) (2) (3) (4) (5) (6) (7) (8) Only Includes Cases That Occurred Before the Year < 2002 <2003 <2004 <2005 < 2006 < 2007 < 2008 < 2001 0.001 0.044 0.082 0.082 0.130 0.163 0.236 0.234 (0.038) (0.045) (0.041)* (0.045)* (0.050)** (0.061)** (0.077)*** (0.086)** Observations 210 231 252 273 294 315 336 357 R-squared 0.99 0.99 0.99 0.99 0.99 0.98 0.98 0.98 Note: * significant at 10%; ** significant at 5%; *** significant at 1%; Robust standard errors in parentheses. Unit of observation is claim frequencies for each state-year. Each column shows the results from a separate diff-in-diff regressions including only the observations that occurred prior to the year in column heading. Only states that passed apology laws prior to 2002 and states that have never passed an apology laws are included. All regressions include state and year dummies and controlling for other law change includes non-economic damage cap, punitive damage cap, law on medical malpractice disclosure, csr_tort and jsl_tort. Table 5. The Impact of Apology Law on Medical Malpractice Claim Frequencies by Severity of Medical Outcomes Dependent Variable: log (claim frequencies) Major Permanent Injury/Death Insignificant Injury "Somewhat" Significant Injury (1) (2) (3) (4) (5) (6) (7) (8) (9) Adopted Apology Law -0.167 -0.182 -0.194 0.118 0.091 0.047 0.270 0.265 0.217 (0.099)* (0.104)* (0.101)* (0.124) (0.124) (0.121) (0.129)** (0.133)* (0.141) Other Law Changea b X X X X X X Other Covariates X X X State Fixed Effects X X X X X X X X X Year Fixed Effects X X X X X X X X X N 255 255 255 255 255 255 255 255 255 R-squared 0.91 0.91 0.92 0.93 0.93 0.93 0.93 0.93 0.94 Note: * significant at 10%; ** significant at 5%; *** significant at 1%. Each column shows the results from a separate diff-in-diff regressions. Standard errors are clustered at the state level. The dependent variables are Log (claim frequencies by severity of outcome) in a state-year. a. Other law change includes non-economic damage cap, punitive damage cap, law on medical malpractice disclosure, csr and jsl tort b. Covariates include Population, % Black, % White, % of population that are 65 or above, and # of Physicians. Table 6. Duration Analysis of the Impact of Apology Law on Time to Resolution By Severity of Medical Outcomes Insignificant Injury (1) Adopted Apology Law State Fixed Effects Year Fixed Effects (2) 1.116 (0.078) X X Other Law Changea Other Covariates N "Somewhat" Significant Injury (3) 1.126 (0.074)* 1.129 (0.075)* X X X X X X b (4) (5) 1.148 (0.094)* X X (6) 1.146 (0.098) 12864 12864 11242 (7) 1.153 (0.098)* X X X X X X X Major Permanent Injury/Death (8) 1.198 (0.076)*** X X (9) 1.195 (0.080)*** X X X X X X X 23945 23945 22586 1.192 (0.079)*** X 26437 26437 25153 Note: * significant at 10%; ** significant at 5%; *** significant at 1%. The unit of observation is per malpractice claim. Numbers reported above are hazard ratios for case resolution. Each column shows the results from a separate duration analysis with Cox hazard model. The dependent variable is time to resolution (year of resolution minus the year of incident). Errors are clustered at the state level. a. Other law change includes non-economic damage cap, punitive damage cap, csr_tort, jsl_tort & law on information disclosure b. Other covariates include allegation nature, patient gender, settlement type, experience of physician Table 7: The Impact of Apology Law on Claim Payouts By Severity of Medical Outcomes Insignificant Injury Baseline Mean $45,019 (1) (2) "Somewhat" Significant Injury Baseline Mean $155,070 (3) Major Permanent Injury/Death Baseline Mean $342,869 (4) (5) (6) (7) (8) (9) -24017 (13432)* -27264 (13564)** -16989 (9538)* -73096 (17334)*** -67644 (21188)*** -55247 (18022)*** Panel A: Dependent Variable-- Payout in Y2000 dollars Adopted Apology Law -431 (4236) 632 (4131) 3132 (3893) Panel B: Dependent Variable--Log Claim Payout Adopted Apology Law 0.096 (0.051)* Other Law Changea 0.097 (0.053)* 0.140 (0.058)** X X b -0.083 (0.090) -0.097 (0.085) -0.08 (0.071) X X -0.171 (0.075)** -0.177 (0.074)** -0.14 (0.067)** X X Other Covariates State Fixed Effects Year Fixed Effects X X X X X X X X X X X X X X X X X X X X X N 13317 13317 11618 24156 24156 22780 26561 26561 25273 Note: * significant at 10%; ** significant at 5%; *** significant at 1%. In Panel A, numbers reported above are payments in Y2000 dollar. Each column shows the result from a separate diff-in-diff regression. The dependent variable is the dollars value claim payout in Panel A and the dependent variable is the log (claim payout) in Panel B. a. Other law change includes non-economic damage cap, punitive damage cap, csr_tort, jsl_tort & law on information closure b. Other covariates include allegation nature, patient gender, patient age, experience of physician and square of experience