The effects of performance evaluations and trust on popular strategies for dealing with uncertainty about the outcome of critical surgery Dr Neil Munro Lecturer in Chinese Politics “Violence towards healthcare professionals: an assessment of causes, frequency, prevention and solutions” University of Chicago Center, Beijing September 17-18, 2013 . 1 Key Issue • Why is there violence towards health care professionals? • Understanding reasons for behaviour is often the key to stopping it. • If widespread, it may point to a crisis in the relationship between the medical profession and the general public. • Outline – Key concepts – Data – Survey item: strategies to improve outcome of critical surgery – Independent variables: performance evaluations & trust – Effects on choice of strategy – Implications for understanding violence in health care settings 2 Key concepts • Agency dilemma: patients are principals, doctors are their agents. How can patients ensure that the doctor offers as far as possible the same treatment as they would administer to themselves if they were the patient? • Exit: a strategy for resolving the dilemma by changing the doctor or health care provider for another one, presumed to be more loyal to the patient’s interests • Voice: a strategy for resolving the dilemma by speaking or acting to influence the behaviour of the doctor or health care provider • Particularism: a strategy for resolving the dilemma which helps a particular patient but not others in a similar situation. – VIOLENCE is a particularistic form of voice-in-action. 3 Table 1. Three Dimensions of Strategy Proactive Passive Universalistic Voice Exit Make a Change formal provider complaint Particularistic Voice Exit Pay a bribe, Use Ask the connections provider to to find change another their provider decision Accept the treatment without resolving the dilemma, Don’t know, Give up treatment Source: adapted from Hirschman, A. O. (1970). Exit, Voice and Loyalty: Responses to Decline in Firms, Organizations and States. Cambridge, MA, Harvard University Press. 4 Why use violence? • Violence implies a step change in degree of conflict. • Violence is qualitatively different from other forms of voice. It is a transgression of universal norms of acceptable social behaviour. • Two kinds of logic can be used to explain it (March & Olsen 1989): – Logic of consequences (egocentric logic): violence will help me get what I want, and consequences for me are acceptable; therefore, I will use it. – Logic of appropriateness (socio-tropic logic): there are rules of conduct in my subculture under which violence is regarded as an appropriate response in certain circumstances. 5 Data • Performance Evaluations, Trust and Utilization of Health Care in China Survey, 2012-13. Funded by the UK Economic and Social Research Council, Grant No. ES/J011487/1. • nationwide survey in 31 provinces on mainland China • fieldwork carried out 1 November 2012 to 17 January 2013 • target population: mainland citizens age 18 to 70 residing for more than 30 days in family dwellings • GPS Assisted Area Sampling Method • Sample of 5,424 dwellings: 3,684 valid interviews • Response rate of 67.9 per cent • More information at http://www.gla.ac.uk/petu Figure 1 Strategies to Improve Chances of Successful Surgery B22. If your friend needed a life-saving operation but the surgeon in a public hospital told them that he could not guarantee that the operation would be successful, what would you advise the patient to do? (One answer). (% replies) Source: Performance Evaluations, Trust and Utilization of Health Care in China Survey, fieldwork 1 November 2012-17 January 2013, N=3680. 7 8 Figure 2a. Perceived Prevalence of Overprescribing C21. Some people say medical ethics is a big problem in this country, but others say such reports are just exaggerated. Judging from your own personal experience, how likely do you think it is that you would encounter the following types of situations in city or county hospitals around here? a) Prescribing medicines not covered by insurance even when effective alternatives covered by insurance are available. Source: Performance Evaluations, Trust and Utilization of Health Care in China Survey, 2012-13. 9 Figure 2b. Perceived Prevalence of Taking Hongbao b). …Taking bribes (“red envelopes” or hongbao) for treatment which has already formally been paid for. Source: Performance Evaluations, Trust and Utilization of Health Care in China Survey, 2012-13. 10 Figure 2c. Perceived Prevalence of Unnecessary Testing c). …Requiring comprehensive check-ups from patients even when the diagnosis is perfectly clear. Source: Performance Evaluations, Trust and Utilization of Health Care in China Survey, 2012-13. So how do people respond to uncertainty? • Dependent variable: which strategy does respondent prefer? • Key independent variables: • Performance evaluations: • health system as a whole • skills & competence of doctors • Trust in health care institutions (global trust) • Perceived likelihood of unethical practices (fidelity) • Comparisons between public & private • Controlling for: demographics, education, urbanization, hukou, self-assessed health, difficulty paying health bills, local government spending per capita 11 12 Figure 3a. Impacts of Odds of Using Voice Impact* Overall satisfaction with health care system Agricultural hukou Local gov’t expenditure per capita * Impact on logged risk ratio relative to reference category, computed by multiplying multinomial logistic regression coefficients by the mean of the independent variables. Impacts not significant at .05 level are not shown. Source: Generalized linear mixed model for data in Figure 1, N=3,506. 13 Figure 3b. Impacts on Odds of Using Particularistic Strategies Impact* Struggles to pay family medical bills Overall satisfaction with health care system Unethical practices likely Education Agricultural hukou * Impact on logged risk ratio relative to reference category, computed by multiplying multinomial logistic regression coefficients by the mean of the independent variables. Impacts not significant at .05 level are not shown. Source: Generalized linear mixed model for data in Figure 1, N=3,506 14 Figure 3c. Impact on Odds of Using Exit Strategies Impact* Struggles to pay family medical bills Age in deciles Unethical Private practices treatlikely ment better Education Agricultural hukou Local gov’t expenditure per capita * Impact on logged risk ratio relative to reference category, computed by multiplying multinomial logistic regression coefficients by the mean of the independent variables. Impacts not significant at .05 level are not shown. Source: Generalized linear mixed model for data in Figure 1, N=3,506. Key Findings (System performance) People who are satisfied with performance of the health care system as a whole, are more likely to: • use voice (ask the doctor to change treatment & make complaints) • use connections to get a referral. (Medical ethics) People who think unethical medical practices are widespread are more likely to: • pay hongbao • go to a private hospital. (Trust) People who trust hospitals are less likely to complain to the hospital administration. (Clinical performance) People who are satisfied with the skills & competence of hospital doctors are more likely to complain. 15 A theory about reasons for violence • Unsuccessful surgery can cause people to lose those they love and on whom they depend. • Patients regard payment, especially informal payment, as creating an obligation on doctors to resolve a health care problem. • Failure to fulfil that obligation is a violation of norms of reciprocity. • Reciprocity is a strong norm in Chinese culture. Violence towards health care professionals is an extreme expression of moral indignation about unfulfilled reciprocal obligations. Evidence is circumstantial: need detailed case studies 16