| BackgroundThere were a total of 4.29 million new tumor cases in China,and more than 2.8 million cancer patients died in 2015.Cancer has become the leading cause of death in China.Up to now,quality of end-of-life care for patients with terminal cancer has always been at a low level.Aggressive treatment,lack of pain relief and effective communication impose poor quality of life of patients with advanced cancer.Most patients suffer from economic poverty,physical poverty and mental poverty,which together lead to terminal poverty.It is urgent to improve the quality of end-of-life care for patients with terminal cancer so that they can walk out of terminal poverty.Compared with the inadequate development of end-of-life in China,end-of-life care have been developed and popularized for many years in most developed countries,such as the United Kingdom,Australia,the United States and Canada.We can learn from their rich experience.The important premise to improve the quality of patient-oriented end-of-life care is to fully understand and guide patients' preferences.It is undoubtedly difficult to fully understand and guide patients' preferences.Firstly,because of the uncertainty of health care outcomes,patients with advanced cancer often find it difficult to choose their next health care treatment.Secondly,there is a serious information asymmetry between doctors and patients in the medical service market.The level of sharing decision-making is low.The professional complexity of medical knowledge and inadequate communication result in patients' inability to fully understand the situation of their own diseases and the advantages and disadvantages of treatment.Thirdly,there is imperfect agency relationship between patients and doctors.And,there is also imperfect agency relationship between patients and their families who usually act as entrusted agent spontaneously in the context of tradition culture,because the former usually knows less than the latter.Finally,patients at the end of life may suffer from cognitive and behavioral disorders that prevent them from expressing their views clearly.In conclusion,uncertainty of health care,asymmetry the medical information and imperfect agent relationship may lead to actual health care solutions that may depend on more than just patient preferences.How to understand and guide preferences of patients with terminal cancer is a major challenge to improve the quality of end-of-life care.The previous researches have limitations.In terms of research methods,the standard and repeatability of answering questions using qualitative interview are poor,which is easy to be affected by interviewees,and the sample size is usually small.Due to the heterogeneity of the included literature in terms of research background,methods and results,the results of systematic review articles are mostly classified and described on the original literature.Empirical studies are often based on data from existing databases or hospital data.Patients,preferences are inferred by the fact patients have received treatment or the actual decision,but the actual decision may not be based on patients' own preferences.These methods cannot comprehensively measure patients'preferences from multiple attributes and dimensions.In addition,they can't solve the problem of preference heterogeneity,which makes the universality of the final research results questioned.In terms of research contents,there is almost no research on the preferences of end-of-life care for patients with advanced cancer using discrete choice experiment in China.There is a lack of understanding of end-of-life care preferences for patients with terminal cancer and how to improve the quality of patient-centered end-of-life care.Therefore,in order to make up for the limitations of the existing research methods and research contents,this study used discrete selection experiment with multi-dimensional attributes and levels to quantify end-of-life care preferences in patients with advanced cancer.This research can be helpful to fill the gap in research in China.ObjectivesThe general objective of this study is to quantify patients' preferences and willingness to pay for end-of-life care,based on the random utility theory and method of discrete choice experiment.It will be helpful to improve the quality of patient-centered hospice care,to encourage health policy makers to develop end-of-life care services close to patients' real preferences,and to explore end-of-life care model with Chinese characteristics.There are specific objectives of this study.The first aim is to systematically summarize the theoretical models and research methods for measuring preferences at home and abroad.The second aim is to get data on the preferences of patients with advanced cancer by the method of discrete choice experiment,which can express declarative preferences through tangible attributes and levels by controlling experimental conditions and eliminating the interference of confounding factors.The third aim is to quantify the preference of patients,and compare the importance of attributes such as hospitalization days,medical expenses,life extension,quality of life,side effect and place of death on preference.The fourth aim is to predict the probability of acceptance under different scenarios.The last aim is to analyze the heterogeneity of preferences of subgroups,such as the sharing decision-making,quality of life,gender,age,education level,income level and other factors affecting preferences.MethodsIn this study,a discrete selection experiment based on random utility theory was used to measure patients' stated preferences.By means of literature review,qualitative interview,and expert consultation,six attributes are selected including inpatient days,medical costs,prolong life,the quality of life,side effect and place of death.STATA 14.0 was used with D-efficiency design.The optimal design of experiment contains 16 choice sets.Each choice set contains two options.16 alternatives by software is divided into two versions.Each respondent completed eight choice tasks.General label items,warm-up questions and utility test questions were used in the questionnaire design,and a series of deviations were avoided or reduced.Cancer patients aged over 50 years old,who are hospitalized and treated with stage 3 or 4 are recruited.Between August and November,2018,four specially trained investigators went to Grade A comprehensive hospital in Shandong province to conduct survey with face-to-face interview using self-made questionnaire.In this study,Orme equation was used to determine the minimum acceptable sample size.Finally,183 samples were obtained.In this study,Stata 14.0 and Microsoft Office Excel 2007 software were used for statistical analysis.Dummy variable encoding was used to encode the data from discrete choice experiment.Conditional logit model and mixed-logit model were used to model.The optimal model was determined by the goodness of fit of the model which was estimated through Akaike Information Criterion,Bayesian Information Criterion and Log likelihood.Finally,the variability of preferences and the heterogeneity of preferences of different groups were analyzed based on the optimal model.The estimated 95%CI and P values of parameters were listed in the regression analysis results.The P values on both sides were less than 0.05,indicating statistical significanceResults(1)In terms of model comparison,the main results were similarly in Conditional logit model and Mixed logit model.But preference weights were different.Through the comparison of AIC,BIC and LL value,we found that the LL and BIC values in the two models were similar.But the AIC value of Mixed logit model is less than that of Conditional logit model(1717.075<1780.822),which means the goodness of fit of Mixed logit model is better than that of Conditional logit model in our study.Therefore,Mixed logit model was selected as the analysis model based on AIC index.Besides,analysis of willingness to pay analysis was estimated using Mixed logit model.(2)In terms of preference,moderate survival,best quality of life,lower side effect,lower out-of-pocket costs,and death at home significantly influence preferences of end-of-life care(p<0.05).According to preference weight,the most important attributes affecting patient preference were survival and quality of life.To be specific,the preference weight for best quality of life was the highest.Compared with the level of poor quality of life,the level of best quality of life can increase the utility by 1.79 times,followed by the general level of quality of life increasing the utility by 1.29 times and the level of good quality of life increasing by 1.13 times.In addition,compared with 4 months,the utility increased by 1.63 times when the life can be prolonged by 10 months,which was higher than that increased by 6-months survival time and 16-months survival time.Besides,compared with no side effect.low side effect occurred no significant influence on preference,but moderate side effect and high side effect can influence patients' preferences significantly.Patients prefer to die at home.However,hospitalization days was not significant factor affecting patients 'preferences.In conclusion,patients with advanced cancer prefer to get best quality of life,moderately survival time,dying at home and avoid the high incidence of adverse reaction at the end of life.(3)In terms of willingness to pay,patients' willingness to pay for survival and quality of life attributes were higher than another attributes.Compared with poor quality of life,patients' willingness to pay for best quality of life was the highest(i.e.the score of quality of life was increased from 4 points to 10 points),as high as RMB256895.45.If survival time was extended from 4 months to 10 months,patients were willing to pay RMB233446.16,higher than that for 16-months survival time.If the incidence of side effect increased to 50%,patient's willingness to pay was RMB-82077.61.If the incidence of side effect increased to 90%,patient's willingness to pay was RMB-71347.96.Besides,patients were willing to pay RMB 58581.93 to die at home.Based on the results of the WTP,we predict patients,willingness to pay for the optimal end-of-life care was RMB548923.54.The optimal care was included with life extension for 10 months,best quality of life,without side effect and death at home.In general,patients were willing to pay much more for better quality of life and moderate survival time than other attributes.The willingness to pay to get highest quality of life was highest at the end of life.(4)Compared with poor quality of life,99.90%of patients preferred the level of good quality of life,higher than the other two levels(71.61%and 76.45%respectively).99.38%preferred life extension of 16 months rather than life extension of 4 months,higher than the other two levels(92.84%and 75.25%respectively).Compared with death in hospital,88.70%of patients prefer to die at home.(5)The low score of shared decision making indicated that cancer patients were underrepresented in treatment decisions.As measured by SDM-Q-9,the average score of shared decision making was 36.05(the score ranges from 0 to 100).48.63%of patients were unaware of the treatment plan and did not participate in the whole decision-making process.Only 15.85%knew their treatment plan very well and participated in the whole decision-making process.Patients with shared decision-making had a higher preference for quality of life attributes(311739.40>139113.10).,and were more willing to avoid adverse reactions(-153123.00<-102647.50),and die at home(106031.40>61929.70).However,Patients with shared decision-making had lower willingness to pay for the 6-month survival time than that of the non-sharing decision group(158184.10<183235.90).In the preferences heterogeneity of different populations,we also find gender,age,education level,income level,quality of life and other factors can affect preferences,but which needs to be further explored.(6)In terms of preferences for participation in decision-making,we found that 27.32%of patients were unwilling to specify treatment decision goals.They entrusted their decisions to a doctor or an agent,so they were "positive" patients.On the opposite,72%of patients directly expressed their goals for treatment decisions,so they were active.In addition,when patients cannot communicate with doctors on their own,they delegated children and spouses was their main agents.Specifically,69.40%of patients were willing to entrust their children to make decision,and 26.78%of patients were willing to entrust their spouses.(7)We found that the actual situation of end-of-life care received by patients was far from meeting the needs of cancer patients through qualitative analysis,especially in terms of quality of life.ConclusionThere is a serious inconsistency between the revealed preference and stated preferences for end-of-life care among patients with terminal cancer.Life extension is not the most important consideration for cancer patients.Patients are more willing to pursue a higher quality of life.Patients prefer end-of-life care with higher quality of life,moderate life extension,less out-of-pocked payment,less side effect and death at home.To be specific,firstly,life extension is not the most important attributes for patients,but best quality of life is.Patients reported the highest WTP to obtain a higher quality of life,and they are not willing to blindly extend life.Secondly,home is a better place to die.Thirdly,medical cost is another important attribute that affects patient's preference.Fourthly,the inpatient days and lower side effect had no significant influence on patients' preference.Fifthly,the degree of shared decision-making between patient and doctor is low.Improving the degree of shared decision-making can help guide patients to prefer a higher quality of life,avoid adverse reactions and death at home.At last,demographic characteristics will affect the preference heterogeneity to some extent.Most cancer patients are willing to entrust their children as agents.Therefore,it is urgent to establish end-of-life care model with Chinese characteristics and high quality.Firstly,it is important and useful to conduct a preference survey for each cancer patient,and encourage the use of advance directives to encourage and guide patients to express their real preferences and decision.Decision makers should pay attention to preferences heterogeneity of different groups.Secondly,taking measures to strengthen pain management is necessary.Thirdly,our government and health providers should take measures to improve the accessibility and the availability of end-of-life care.Fourthly,we should encourage and strengthen end-of-life care education and training in various forms,strengthen training on communication skills,and improve the degree of shared decision-making.Fifthly,it is suggested to set up a hospice medical insurance program,including end-of-life care into the reimbursement of medical insurance.It is useful to reduce the proportion of out-of-pocket expenses for patients at the end of life.In addition,we should establish special charge standards for hospice care wards.End of life care should be integrated into the social security system.In the end,we will strengthen government intervention with formulating relevant laws,policies and regulations,and formulating sound evaluation standards,which can provide legal support for the establishment of end-of-life care system with laws to abide by. |
PDF Full Text Request