How common are Chinese patients with multimorbidity involved in decision‐making and having a treatment plan? A cross‐sectional study

Creating a treatment plan (TP) through shared decision‐making (SDM) with healthcare professionals is of paramount importance for patients with multimorbidity (MM). This study aims to estimate the prevalence of SDM and TP in patients with MM and study the association between SDM/TP with patients' confidence to manage their diseases and hospitalization within the previous 1 year.

and conflicting lifestyle advice and prescription of medications. [7][8][9][10][11][12][13][14] Therefore, instead of following clinical guidelines, the American Geriatric Society 15 and The National Institute for Health and Care Excellence (NICE) 16 recommended a shared decision process to individualize a treatment plan (TP), which is in accordance with patients' preferences and values, and this may minimize treatment burden and maximize quality of life. Shared decision-making (SDM) is defined as "an approach in which the clinician and patient go through all phases of the decision-making process together and in which they share the preference for treatment and reach an agreement on treatment choice"; this is in contrast to the traditional medical model where doctors are solely responsible for prescribing the "best" treatment to patients. 17 Creating TPs by SDM has been shown to enhance patients' sense of control over their illness, improve their symptoms, enhance their knowledge and reduce concerns towards illnesses; which in turn can enhance adherence to medications and improve quality of life. 17,18 Although there were previous studies in Hong Kong investigating SDM in other patient populations, including donot-resuscitate decisions in patients with chronic obstructive lung diseases and in surgical and medical patients, 19,20 the prevalence of SDM and TP in patients with MM, who could benefit most from SDM in Hong Kong, was not previously known. Similarly, it remains unclear if the presence of SDM and/or TP can actually improve patients' outcomes such as a reduction in hospitalization.
The primary aim of this study was to determine the proportion of patients with MM who reported having SDM and/or a TP. As a secondary objective, participants were asked how confident they were to manage their illnesses and whether they have been hospitalized overnight in the past 1 year. The relationships between the presence of SDM and/or a TP and their confidence level to manage their illnesses and history of hospitalization were delineated. We hypothesized that a high proportion of patients with MM had SDM and/ or a TP and having these could enhance their confidence in disease management and reduce hospitalization.

| ME THOD
This cross-sectional study used an internationally recognized survey (see below). A total of 1032 patients were recruited, who (i) were aged 60 or above and (ii) who self-reported to have at least three chronic conditions (Appendix 1). Patients were recruited at one general outpatient clinic (GOPC), one geriatric specialist outpatient clinic (SOPC) and/or geriatric day hospital (GDH) in each of the seven HA clusters, from June 2016 to July 2017. Written consent was obtained from all participants before their participation of the project.
We included patients only older than 60 years because MM is most common in the older population. Besides, all patients in GDH and SOPC were older than 60 years. While Hong Kong has a dual healthcare system where patients can choose to obtain healthcare from both private and the public sector, the vast majority of patients with chronic diseases were seen under the Hospital Authority system, where the current study was conducted. 21 The questions used in the current research were extracted from the International Health Policy Survey of Older Adults, which was used previously in multinational research involving 11 countries and more than 15 000 participants. 22 The instrument consists of questions to estimate or understand healthcare costs and access, doctor-patient relationships, health promotion, management of chronic conditions and caregiving. 22 The survey has been used in various other large-scale research projects. 23 The survey was translated and validated by iterative forward-backward translation and cognitive debriefing interviews in Hong Kong. 24 After the questionnaire was piloted, a few questions were added by the expert panel, which consisted of three clinical and social experts involved in the care of elderly people, each possessing at least 10 years of working experience. These additional questions were aimed to fine-tune the instrument to fit the local cultural context. Demographics including number of chronic illnesses, age, sex, education level, marital status, financial income sources [social allowance Comprehensive Social Security Assistance (CSSA) signifies disadvantages financially], regular healthcare provider for participants' chronic illness, whether the participant had health insurance, frailty level and number of medications were collected. Frailty level was measured by the validated Chinese version of five-item FRAIL scale; phenotypes of being robust, prefrail and frail were represented by score 0, 1-2 and 3-5 in the scale, respectively. 25,26 Questions about (i) whether patients had SDM and TP, (ii) patient's confidence in managing their chronic conditions and (iii) patient's number of hospitalizations in the past 1 year were extracted to investigate both their mutual associations and their associations with aforementioned demographic characteristics.

| S TATIS TIC AL ANALYS IS
The demographic characteristics of the study participants were summarized as count and percentage. The outcomes collected by 4-option items were simplified into two levels to facilitate their analysis and interpretation of the results. The proportion of patients who were involved in components of SDM and TP, as well as confidence in managing their chronic conditions and hospitalization history, was presented. Logistic regression was constructed to study the relationship between various demographic data and the presence of SDM process and TP. Variables, set at P-value <.1 in the initial univariate analyses, were entered into the multivariate model to determine the most significant associations. The associations between the presence of shared decision process and TP and patients' confidence in managing chronic disease and history of hospitalization, were also studied using logistic regression.
Results were adjusted for demographics data and confounding factors in model 1 and were further adjusted for mutual effects of SDM and TP in model 2. Because the complexity of MM depends on the nature of the diseases and their combination, sensitivity analysis to detect effect of individual disease on our outcomes was conducted using the Jackknife method, 27 which replicates the main analyses multiple times with patients having each of the chronic conditions used to define MM excluded. 28 Odds ratio (OR) and 95% confidence interval (CI) were estimated to clarify the strength of association, and the significance is considered a two-sided P < .05. Statistical analyses were conducted using IBM SPSS Statistics 21.
The percentages of missing data for four major outcomes were lower that 6%. Missing data were assumed to be missing at random; therefore, our analysis was based on complete cases only. No characteristics differences were found between subjects with and without missing values.

| SAMPLE S IZE
Because of a lack of previous similar studies, at a precision of 3.1% and a presumed prevalence of 50% of patients with MM who received SDM (which required the largest possible sample size), the required sample was determined to be 1000 participants. Therefore, our sample size of 1032 patients was considered adequate.

| Participants
Out of 2331 patients that were approached, 1032 patients completed the questionnaire and the response rate was, therefore, 44.3%. More than one third of our participants were older than 80 years (35.2%). Around half of them were male (53.5%), and most participants received some level of education (80.9%), were married (67.4%), had three to four diseases (60.9%), received ≥4 medications (52.3%) and had no health insurance (90.9%). Around one-fifth were classified as being "robust" (21.9%) and were receiving CSSA (16.7%).
The vast majority reported that they had regular healthcare facilities to visit or regular doctors (93.1%). (Table 1).

| Proportion of the presence of shared decisionmaking and a treatment plan
Of patients, 82.1% reported the presence of TP. But for an SDM process, only those having follow-up by regular doctor/facility were guided to answer this item, and among them, only 35.8% of patients reported an SDM process ( Table 2).

| Factors associated with the presence of shared decision-making and a treatment plan
The presence of the SDM process was not associated with any demographic data, including age, sex, education, marital status, .980 a The response of "Sometimes/rarely/never" was considered as reference category. b No treatment plan was considered as reference category.
c Only patients answered "Yes" in the item of "Regular doctor/facility" were guided to answer the item "Deciding a treatment according to your will and get you involved." *P < .05; **P  Our sensitivity analysis by the Jack-knife approach showed that regular doctor/facility remained an important factor for the presence of TP in most models (Appendix 2). However, participants being recruited from SOPC and/or GDH, who were less likely to receive a TP when chronic pain conditions, eye diseases, diabetes mellitus and musculoskeletal diseases, were excluded.

| Association between the presence of shared decision-making/treatment plan and patients' confidence to manage disease and hospitalization
Overall, a quarter of patients (25%) felt not confident enough to manage their health problems. In the fully adjusted model, the presence of TP enhanced patients' confidence to manage their diseases (OR = 2.384, P < .001) (

| D ISCUSS I ON
This is one of the first studies that explores the prevalence of SDM and TP in Chinese patients with MM, which showed that the presence of SDM and/or TP was associated with enhanced patients' confidence to manage their illnesses. Previous similar studies involved Chinese patients with breast cancer and found inconclusive results.
One study revealed that 70% of patients were allowed to decide their preferred surgery 29 ; but a second study mentioned that the level of SDM in which these patients were engaged was low, according to a validated scale using direct observations of the actual consultations. 30 In the current study, the majority (82.1%) of participants were aware of a TP, but only around one-third of participants recalled having an SDM process in which their priorities, and preferences were taken into consideration to build the TP. It was likely that TPs were prescribed by doctors rather than as a product of discussion Note: Multivariate models include variables which P < .10 in the univariate analysis. The item "Deciding a treatment according to your will and get you involved" was not shown here due to all variables had P ≥ .10 in the univariate analysis. a No treatment plan was considered as reference category; *P < .05.
are more likely to formulate a TP that patients can recall. 31 However, the presence of SDM or TP was not associated with the history of hospitalization, suggesting that hospitalization was mainly driven by the progression of diseases and actual needs. In fact, when patients with hypertension or diabetes were excluded, TP was associated with increased hospitalization. It is possible that TP included advices to observe for alarming symptoms or signs, and this may prompt hospitalization when physical conditions deteriorate.
SDM was reported only infrequently in our sample, despite its internationally recognized importance in patients with MM. 15,16 The prevalence of SDM in MM in other countries was underreported, and the current study is one of the first that reported the prevalence of SDM and TD in patients with MM. However, SDM remained underutilized in many populations (eg, without MM) internationally; for example, a study found that around only half of the seriously ill patients who wished to refuse resuscitation had a "do-not-resuscitate" order, and healthcare professionals were found to have a poor understanding of these preferences. 32 Yet, SDM might improve patients' outcomes. A Cochrane review of randomized controlled trials supported that the involvement of patients using decisional aids could improve their knowledge and reduce internal conflicts within decision-making. 33 A cohort study in women with breast cancer also suggested that SDM enhanced patients' quality of life. 34 Evidence also suggested that SDM may reduce the financial burden of healthcare systems because when provided with choices, participants often opted for more conservative, rather than intensive and expensive treatments. 33  While decisional aids were suggested to help patients make informed decisions, the relevance of these aids to patients with MM was uncertain because these decisional aids were usually diseasefocused and were only available for a limited spectrum of diseases. 36 The use of decisional aids in Chinese contexts is especially understudied. 33 Doctors can be reluctant to use decisional aids during d "Not very confident/not at all" was considered as reference category. A total of 255 patients answered "Not very confident/not all" (25%), whereas 764 patients answered "Very confident/confident" (75%) in this item. e" No" was considered as reference category. A total of 609 patients gave response of "No" (59.5%), whereas 414 patients gave response of "Yes" in this item (40.5%).
* P < .01. consultation because they can lengthen the consultation time by 2.6 minutes, 33 whereas the average consultation time in GOPC is around 5-7 minutes in Hong Kong. 37 Furthermore, many patients, especially Chinese, may not want to be involved in the decisionmaking process 38 ; doctors may be reluctant to involve patients in making decisions if they perceive patients to be unwilling to make a decision, or if the patients were not educated enough to engage in such a discussion. 35 Previous research showed that older Chinese people are less willing to make health-related decisions and the presence of SDM depended also on patients' education level. 29,30 The latest Cochrane Review also suggested that there is a lack of evidence of ways to encourage clinicians to involve patients in making decisions. 39 Research on interventions to promote SDM in our patients with MM is, therefore, needed; such trials can then provide evidence on health benefits and cost-effectiveness, if any, of SDM.
The current study recruited more than 1000 patients with MM from both primary and specialist clinics from all areas of Hong Kong and represented one of the largest studies in a Chinese population.
However, a few limitations must be discussed.

| CON CLUS ION
In conclusion, most patients with MM in Hong Kong had a TP, but fewer had been involved in making health-related decisions. TPs and shared decision processes appeared to help patients to manage their diseases. Going forward, research is needed on interventions that promote SDM in patients with MM.

ACK N OWLED G EM ENTS
Sincere thanks to all involved Hospital Authority staff for their logistic supports; and also thanks to all participants for their participation in this study.

CO N FLI C T O F I NTE R E S T
The authors declare that they have no competing interests.

AUTH O R CO NTR I B UTI O N S
KPL, SYSW, BHKY, ELYW, DC and EKY were responsible for the literature review section. They also contributed in creating and organizing the figures, as well as the design for the above study. In addition, they were involved in data analysis, data interpretation and writing the manuscript. PC and LL were also involved in the data collection and data analysis. All authors read and approved the final version of the manuscript.

DATA AVA I L A B I L I T Y S TAT E M E N T
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request. Note: Only chronic diseases with cases of more than one third of the sample size were included in this table.

E TH I C S A PPROVA L A N D CO N S E NT TO PA RTI CI PATE
a Not very confident/not at all was considered as reference category. b No hospitalized overnight during the past 1 year was considered as reference category.
c Adjusted by age group, education and type of clinic; and ORs of SDM and TP were mutually adjusted.