Healthcare needs, experiences and treatment burden in primary care patients with multimorbidity: An evaluation of process of care from patients' perspectives

Abstract Background Patients with multimorbidity often experience treatment burden as a result of fragmented, specialist‐driven healthcare. The ‘family doctor team' is an emerging service model in China to address the increasing need for high‐quality routine primary care. Objective This study aimed to explore the extent to which treatment burden was associated with healthcare needs and patients' experiences. Methods Multisite surveys were conducted in primary care facilities in Guangdong province, southern China. Interviewer‐administered questionnaires were used to collect data from patients (N = 2160) who had ≥2 clinically diagnosed long‐term conditions (multimorbidity) and had ≥1 clinical encounter in the past 12 months since enrolment registration with the family doctor team. Patients' experiences and treatment burden were measured using a previously validated Chinese version of the Primary Care Assessment Tool (PCAT) and the Treatment Burden Questionnaire, respectively. Results The mean age of the patients was 61.4 years, and slightly over half were females. Patients who had a family doctor team as the primary source of care reported significantly higher PCAT scores (mean difference 7.2 points, p < .001) and lower treatment burden scores (mean difference −6.4 points, p < .001) when compared to those who often bypassed primary care. Greater healthcare needs were significantly correlated with increased treatment burden (β‐coefficient 1.965, p < .001), whilst better patients' experiences were associated with lower treatment burden (β‐coefficient −0.252, p < .001) after adjusting for confounders. Conclusion The inverse association between patients' experiences and treatment burden supports the importance of primary care in managing patients with multimorbidity. Patient Contribution Primary care service users were involved in the instrument development and data collection.


| INTRODUCTION
Multimorbidity-the presence of two or more chronic conditions within an individual-has become increasingly common over recent decades. [1][2][3][4] It presents complex challenges to patients, such as functional decline, mental health difficulties, polypharmacy, reduced quality of life, increased hospital admission and risk of severe COVID-19. [5][6][7][8][9][10] Existing evidence supports the role of high-quality primary care in improving population health outcomes in a costeffective manner, and primary care is of particular importance in addressing multiple healthcare needs. 1,11 China, like many countries that are facing health and social care challenges from an ageing population, is reshaping its healthcare system with a primary care-oriented approach to pursue equitable population health and reduce the burden of chronic conditions. 12,13 Primary care facilities have been established for delivering safe, effective, convenient and affordable healthcare by general practice (GP) physicians outside of hospitals. However, healthcare gatekeeping is largely absent and thus people can bypass primary care and go straight to hospitals for specialist care as they wish. The concept of a 'family doctor team' has been gradually translated into practice since June 2016 as an emerging healthcare model built on the national basic public health (BPH) service package. 14,15 A typical team is comprised of one GP clinician and several healthcare personnel including nurses, public health doctors and, if available and suitable, pharmacists and social workers. This supports a broader range of systematic preventive care approaches, including health assessment, health promoting interventions, health advice and, when necessary, home visits to support self-management. The primary care multidisciplinary team is expected to be responsible for the health of enrolled people and their family members. 15,16 The management of a population with multimorbidity requires routine primary care that is respectful of, and responsive to, their increasing need for family-centred continuity of care as opposed to hospital-based fragmented care. In a fragmented healthcare system, it is less likely that multiple, episodic healthcare providers will take into account the entirety of a patient's healthcare conundrum including inappropriate polypharmacy, demanding self-management regimens and competing priorities and more vulnerability to safety issues due to multimorbidity. 3 This would inevitably lead to unaddressed issues associated with greater lapses in quality and safety, higher healthcare expenditure and more avoidable hospital admission. 9,17 Management of multimorbidity is complex and necessitates coping strategies built upon continuous care with consultations, examinations, medications and lifestyle changes, placing significant burden on patients in terms of excessive time, efforts and attention. 4,[18][19][20] The understanding of processes of care that take into account patients' healthcare needs and minimize treatment burden is an essential step to inform service delivery for multimorbidity. 21 These relationships between need, patients' experiences, treatment burden and use of primary care have not been described before, and are highly relevant in the context of the growing challenge of multimorbidity globally.
This study aimed to provide an insight into healthcare needs, patients' experiences and treatment burden from the perspective of process of care. Our key research question is whether there is a significant association between primary care experiences and treatment burden in the context of patients' increasing healthcare needs due to multimorbidity. In the absence of a secondary healthcare gatekeeping function in primary care, we hypothesize that patients who do not consider the family doctor team as their preferred usual source of care will have poorer primary care experiences and greater treatment burden.

| Study design
Multisite cross-sectional survey data were collected from primary care service users, with a diversity of geographic locations, in 9 out of a total of 21 cities in Guangdong province, southern China. In the first stage, three cities were selected in each of the western, central and eastern areas of Guangdong, respectively. In the second stage, two sites per city were randomly selected from primary care facilities that were organizational members of the Guangdong Primary Healthcare Association to facilitate the fieldwork coordination.

| Setting and data source
The study was conducted on-site at 18 primary care facilities where free-of-charge, annual check-up, as part of the national BPH service package, was offered to people aged ≥35 years who had hypertension or diabetes. 14 Routine primary care patients who fulfilled the eligibility criteria were invited on the day of their check-up visits at community health centres (CHCs). Our previous work showed that a minimum of 2500 community residents had enrolment registration with the CHC family doctor team, 12 and that more than 10% of the general population had ≥2 chronic conditions (multimorbidity). 2 We assumed a check-up attendance rate of at least 60% and a survey response rate of no less than 80%. This yielded a sample size of 2160-that is, 120 participants recruited in each CHC. Intervieweradministered questionnaires including items derived from our previous research 2,22 were used to collect data on demographics, socioeconomic status, health characteristics, healthcare needs, service utilization and the process of care from study participants.

| Participants
The inclusion criteria of target participants were as follows: (1) patients who had ≥2 clinically diagnosed long-term conditions including hypertension or type 2 diabetes and (2) had at least one clinical encounter in the past 12 months since enrolment registration with the family doctor team. We excluded those who were passers-by (i.e., patients who were not enrolled or only recently enrolled with the family doctor team) to ensure that all study participants had valid exposure to the primary care provider before study participation, and could hence minimize the likelihood of capturing 'hearsay' information that was not actually experienced by the patients. Patients who were unable to communicate or who were not on regular medications were excluded. Eligible participants were referred to trained interviewers by healthcare staff, following a modified systematic random sampling that was previously used. 22

| Measurements of patients' experiences
Patients' experiences were captured by a previously validated, culturally adapted, Mandarin Chinese version of the Primary Care Assessment Tool (PCAT)-Adult Edition used in our previous research. 13,22 The instrument measures nine primary care attributes, that is, the first-contact accessibility and utilization (first-contact domain), continuity of care (longitudinal domain), coordination of services and information system (coordination domain), comprehensiveness of service availability and provision (comprehensiveness domain) and community orientation and family centredness (derivative domain). 23 First-contact care accessibility refers to whether patients are able to receive primary care whenever needed within a reasonable time in nonemergency situations, whereas first-contact care utilization measures the extent to which a gatekeeper function is performed by the primary care provider. Coordination of care services assesses the linkage of healthcare visits across different levels in the health system, whereas the information system coordination measures the availability of health records for patients. All individual items were scored on a 4-point Likert-type scale, with higher scores indicating more positive experiences. 22,23 The total PCAT scores were calculated by summing up values from each of the nine scales.
An adapted algorithm from the PCAT guideline was used to identify respondents' usual source of care, including both frequent and less frequent primary care service users. 22

| Measurements of treatment burden
Treatment burden was defined as the challenges that patients face in coping with everything they have to do to take care of their health, and its impact on functioning and well-being. 20,[24][25][26] It involves a variety of treatment workloads pertaining to medication management, self-monitoring, laboratory tests, doctor visits, need for organization, administrative tasks, lifestyle changes and social impact. 27,28 The Treatment Burden Questionnaire (TBQ) is one of several existing measures, but was specifically developed to assess treatment burden among patients with multiple chronic conditions. 28 It is composed of 15 items using a 10-point rating scale, ranging from 0 (not a problem) to 10 (big problem). The sum of all item scores was calculated, and higher scores indicated greater treatment burden. 28 A total TBQ score of 59 is a recommended cutoff for defining high burden. 26   Training sessions were held by the two lead investigators. The survey was pilot-tested by paired interviewers among 20 primary care service users to improve the interrater reliability.

| Statistical analysis
Data entry was independently performed by two trained medical students using EpiData 3.1 with double verification. Sample mean with standard error (SE) or 95% confidence interval (CI), where appropriate, was applied in descriptive analysis. χ 2 tests or Student's t tests, where appropriate, were used to compare the differences with regard to categorical and continuous variables between groups. A general linear model analysis was conducted to examine patient-level factors associated with treatment burden after controlling for confounders. The absence of multicollinearity and plausible interactions among variables were tested to ensure the robustness of the linear regression model. We also performed a series of sensitivity analyses to further explore the relationship between primary care experiences and each treatment burden measure, while controlling for other confounding factors in the multiple linear regression analysis. A p-value <.05 was considered statistically significant. All statistical analyses were performed in IBM SPSS Statistics 25, and the Complex Samples module was used to account for the multistage sample design.

| Ethics consideration
All study participants provided written consent. Data anonymization was performed by removing all patient identifiers from the data set before data analysis. Ethics approval was granted from the School of

Public Health Biomedical Research Ethics Review Committee at Sun
Yat-Sen University (SYSU-SPH2016027) in accordance with the Declaration of Helsinki 2013.

| Characteristics of the study participants
A total of 2160 out of 2471 eligible primary care patients with multimorbidity were included (overall response rate 87.4%). The mean age of the participants was 61.4 years (95% CI: 60.7-62.2 years), and slightly over half were females. Less than half had completed secondary school education or above. Nearly one in five patients had multimorbidity for over 10 years. Approximately 40% of people had a monthly household income per capita below ¥2000.
When compared to China's median disposable personal income (¥2028 per month) in 2018, 29 the study participants were relatively wealthier than the general population (Table 1).

| Profile on service utilization and healthcare needs
More than two thirds (70.1%) of the participants considered the CHC family doctor team as their usual source of primary care. The traditional face-to-face visit was more common than distance  Figure 1).

| Treatment burden and patients' experiences with primary care
Participants reported an average global treatment burden score of 43.9 (SE: 0.9), which was slightly higher than the first quantile of the score range (0-150), whilst a total primary care assessment score of 100.7 (SE: 0.9) was reported on average, falling within the third quantile of the score range (0-132). This implied moderate-to-light treatment burden and medium-to-optimal primary care experiences overall ( Table 2). Significant differences existed across most of the individual primary care scales between groups. In particular, patients who considered the family doctor team as the primary source of care had significantly better patients' experiences (mean difference 7.2 points, 95% CI: 4.6-9.8, p < .001) and lower treatment burden (mean difference −6.4 points, 95% CI: −9.6 to −3.1, p < .001) when compared to their counterparts who were in favour of using specialist care over primary care (Figure 2).  (Table 3). Moreover, in the sensitivity analysis with each individual TBQ item score as the dependent variable in the regression analysis, the negative associations between primary care experiences and treatment burden were consistently observed, except for medicationrelated item scores, albeit that the lower boundary of the 95% CI for the β (PCAT total score) remained negative (Figure 3).

| Strengths and limitations
We collected data from a relatively large sample of Chinese primary care service users with multimorbidity to understand the process of care using widely used international instruments with appropriate linguistic and psychometric validation. A focus on patients' experiences and process-related treatment burden, rather than patients'  9 it is reasonable to assume that frequent users of primary care tend to be more prevalent in less affluent areas, where the strength of associations between patients' experiences and treatment burden might be stronger.

| Comparison with the existing literature
Empirical evidence from low-and middle-income countries suggests that one of the worst-performing areas in primary care is the prevention and management of chronic diseases. 36 Key problems F I G U R E 3 Association between the overall primary care assessment (independent variable) and each treatment burden measure (dependent variable) in the multiple linear regression analysis. Note: PCAT, Primary Care Assessment Tool; TBQ, Treatment Burden Questionnaire. Error bars indicate 95% confidence intervals of the β-coefficients for the PCAT total score (i.e., independent variable; X) in each regression model with regard to each individual item score in the TBQ (i.e., dependent variable; Y), respectively, while controlling for other confounding factors that were statistically significant in the general linear model analysis shown in Table 3. *p < .05; **p < .01; ***p < .001 commonly experienced by multimorbid patients included a lack of holistic care, poor service experiences and a high burden of disease treatment. 37 This has also raised challenges in high-income settings where previous work reported that unfavourable patients' experiences with primary care physicians were associated with a higher risk of hospitalization. 38 This calls for a deserved attention to positive user experiences and competent care emphasizing healthcare needs and individual preferences, given the complexity of multimorbidity.
The PCAT instrument measuring patients' experiences has been widely used 22,[39][40][41][42] ; however, most of these studies have assessed the process performance among service users overall, and research with a specific focus on attributes of primary care is lacking in the multimorbidity context. An understanding of process-based measures as performance indicators is therefore of importance to inform areas for quality improvements in patient-centred care.
As many clinical practice guidelines tend to focus on single conditions, the treatment burden was often assessed only as a subscale of specific disease scales or was considered only for the regimen associated with a particular condition. 20,26,27 Despite varying approaches in the measurements, existing studies consistently reveal that higher levels of treatment burden relate to multimorbidity, access barriers, fragmented care and patient-provider discordance. 43,44 Since Since 2018, other similar tools have emerged for measuring multimorbidity-related treatment burden, such as the MTBQ questionnaire, which was originally developed in elderly patients and may show the complexity of treatment burden from a more multimorbidity-specific angle. 30 However, it is worth noting that a comparison between different treatment burden measurements per se was not the aim of our study. Instead, we are more interested in determining whether primary care experiences were associated with treatment burden in our study population, which consisted of patients with two or more long-term conditions. Given the proven ability of the TBQ to capture a comprehensive dimensionality of treatment burden, we believe that the relationship between patients' experiences and treatment burden observed in our study shall remain largely unchanged regardless of the instrument per se.
In our study, we found that longer duration of chronic diseases, greater healthcare needs and variables pertaining to the process of care, such as inadequate follow-up and suboptimal primary care experiences, were associated with increased treatment burden. This could be explained by the speculation that patients' greater healthcare needs and extra efforts required to maintain their health may translate into additional workload, such as greater use of medications and challenges in behavioural modifications in coping with multimorbidity. Recent evidence suggests that across a wide range of health conditions and settings, a significant proportion of treatment burden results from the way in which healthcare is organized and delivered, rather than by specific patients, diseases or treatments. 20 Our results confirmed that using primary care regularly and receiving frequently delivered follow-up care with better patients' experiences correlated with alleviated treatment burden, which may imply a positive role of the primary care multidisciplinary team in the process of service delivery. This will also help contribute to the understanding of the extent to which routine interactions between patients and healthcare providers may impact on challenges that patients face in coping with multimorbidity in terms of processes of care.
Further, from a quantitative perspective, our sensitive analysis revealed that a higher primary care experience was consistently Patients in our study reported higher primary care assessment scores and lower treatment burden scores when directly compared to previous studies, 26,27,39,40 which may be due to the differences in study participants and settings where the availability, accessibility and acceptability of resources for primary care may differ. It might also be a reflection of potential gains from an improved process of care with the 'family doctor team' that aims to translate key attributes of primary care into routine clinical practice. 16 As an emerging service model of multidisciplinary It has been widely recognized that treatment burden prevents optimal adherence to the provision and management of care for people with long-term conditions, 21,35 and thus may reduce the overall effectiveness of the health system. A deeper understanding of the complexities of care experiences and the manifestations of treatment burden will help inform an integrated approach at both practice and policy levels to improved chronic care and service delivery in primary care. 49 Our data suggested that an optimal profile of patients' experiences and lower treatment burden were more likely to be seen in multimorbid patients who used the family doctor team as the usual source of primary care. This may lay the foundation for future work to explore the long-term benefits of improved adherence to clinical recommendations and the potential impact of active participation in multimorbidity care plans on desired health outcomes.

| CONCLUSION
Our study suggested that higher healthcare needs were significantly associated with increased treatment burden, whilst better patients' experiences were associated with lower treatment burden in the context of the family doctor team service delivery. This implies the necessity of optimizing the key attributes of primary care in personcentred service delivery and quality improvement, and is therefore of major relevance to healthcare strategies aiming to deliver less burdensome care for people with multimorbidity.

ACKNOWLEDGEMENTS
We wish to acknowledge the tremendous support of the Guangdong-

CONFLICT OF INTERESTS
The authors declare that there are no conflict of interests.