International Journal of Cardiology 177 (2014) 477–482
Contents lists available at ScienceDirect
International Journal of Cardiology journal homepage: www.elsevier.com/locate/ijcard
The association between multimorbidity and poor adherence with cardiovascular medications☆,☆☆ Martin C.S. Wong a,1, Jing Liu b,1, Shenglai Zhou c, Shiwei Li d, Xuefen Su a, Harry H.X. Wang a, Roger Y.N. Chung a, Benjamin H.K. Yip a, Samuel Y.S. Wong a, Joseph T.F. Lau a,⁎ a
School of Public Health and Primary Care, Faculty of Medicine, Chinese University of Hong Kong, Hong Kong Research Centre for Healthcare Management, School of Economics and Management, Tsinghua University, Beijing, China Beijing Anzhen Hospital, Beijing, China d Health Bureau of Weidong District, Pingdingshan, Henan Province, China b c
a r t i c l e
i n f o
Article history: Received 5 August 2014 Received in revised form 17 September 2014 Accepted 20 September 2014 Available online 28 September 2014 Keywords: Multimorbidity Medication adherence Associated factors Cardiovascular medications
a b s t r a c t Multimorbidity, defined as the presence of two or more chronic conditions, leads to a substantial public health burden. This study evaluated its association with adherence with cardiovascular medications in a Chinese population. A proportional stratified sampling was adopted to draw a representative sample of residents living in Henan Province, China. Interviewer-administered surveys were conducted by trained researchers. The outcomes included the number of chronic medical conditions, adherence with long-term medications (MMAS-8), and depressive symptoms (CESD-20). Binary logistic regression analysis was conducted to evaluate if medication adherence was associated with the presence of multimorbidity. From a total of 3866 completed surveys, the proportion of subjects having 0, 1 and ≥2 chronic conditions was 62.6%, 23.8% and 13.5%, respectively. Among 27.6% who were taking chronic medications, 66.6% had poor medication adherence (MMAS-8 score ≤ 6). From binary logistic regression analysis, subjects with poor medication adherence were significantly associated with multimorbidity (adjusted odds ratio [AOR]: 1.35, 95% C.I. 1.02–1.78, p = 0.037). Other associated factors included older age (AOR = 1.04, 95% C.I. 1.03–1.05, p b 0.001), smoking (AOR = 1.63, 95% C.I. 1.16–2.30, p = 0.005), family history of hypertension (AOR = 1.51, 95% C.I. 1.19–1.93, p = 0.001), and fair to poor self-perceived health status (AOR = 2.15, 95% C.I. 1.69–2.74, p b 0.001). Using medication adherence as the outcome variable, multimorbidity was significantly associated with poor drug adherence (AOR = 1.34, 95% C.I. 1.02–1.77, p = 0.037). Multimorbidity was associated with poorer medication adherence. This implies the need for closer monitoring of the medication taking behavior among those with multiple chronic conditions. © 2014 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Multimorbidity is defined as the coexistence of two or more chronic diseases in an individual, and has been increasingly identified as a usual phenomenon in patients [1]. A recent epidemiology study in a database of 1.75 million patients registered at 314 practices in Scotland reported that up to 23% of all people suffered from multimorbidity [1], rising sharply with age. It presents serious challenges to patients and physicians, requiring a patient-oriented approach supported by close ☆ Conflict of interest: None declared. ☆☆ All authors take responsibility for all aspects of the reliability and freedom from bias of the data presented and their discussed interpretation. ⁎ Corresponding author at: 5/F, School of Public Health and Primary Care, Prince of Wales Hospital, Shatin, New Territories, Hong Kong. Tel.: +852 2252 8727; fax: +852 2645 3098. E-mail address: [email protected] (J.T.F. Lau). 1 Co-first authors: The first two authors contributed equally to this manuscript.
http://dx.doi.org/10.1016/j.ijcard.2014.09.103 0167-5273/© 2014 Elsevier Ireland Ltd. All rights reserved.
coordination between generalists and specialists [2,3]. Patients with multimorbidity had more complicated healthcare needs [4], incur higher healthcare costs [5], utilize more health services [6], and were associated with poor clinical outcomes [7]. The recent decade has witnessed a rising incidence of cardiovascular disease (CVD) and its risk factors, including hypertension, diabetes and dyslipidemia [8–10]. They tend to cluster with one another — and arguably patients with a single disease may easily develop into multiple conditions [11]. Persistent adherence with chronic medications according to the healthcare providers' instructions is essential to achieve optimal disease control, and it has been shown that high adherence with cardiovascular medication was associated with a 38% decreased risk of CVD events when compared with lower adherence [12]. It is currently unknown whether patients with multimorbidity may be at higher risks for poor medication adherence. Recent studies mainly focused on the impact of multimorbidity on health-related outcomes
478
M.C.S. Wong et al. / International Journal of Cardiology 177 (2014) 477–482
and health service utilization. As highlighted in some recent reviews, multimorbidity has been regarded as a major research priority [13], yet the number and diversity of evaluations on multimorbidity have been insufficient to provide robust scientific evidence in guiding patient care [14]. The WHO called for action to prioritize healthcare resources for global management of chronic diseases in the next decade, and there is a universal consensus that the topic of multimorbidity should be a central focus of attention [15,16]. We have previously conducted studies on multimorbidity [17–19] and medication adherence [20–25] among hypertensive patients, yet most of them either relied on retrospective data [17,21–24]; or were small-scaled [19,20]. The primary objective of this study is to evaluate the factors associated with multimorbidity and medication adherence in a large Chinese population. Specifically, we tested the a priori hypothesis that the presence of multimorbidity was associated with poorer medication adherence. 2. Methods 2.1. Ethics statement The ethics clearance of the study was obtained from the Survey and Behavioral Research Ethics Committee of The Chinese University of Hong Kong. Informed consent was obtained prior to the study, and each participant was assured of survey confidentiality and anonymity. The study protocol conforms to the ethical guidelines of the 1975 Declaration of Helsinki. 2.2. Sampling frame and methods This study invited participants from the resident health registry of Henan Province, China. The registry consisted of 227,021 residents aged 18 years or older, representing 92.3% of the county's entire population. All the registered individuals were classified into 12 strata, including two strata according to sex (male vs. female), and six strata according to age groups (18–24 years; 25–34 years; 35–44 years; 45–54 years; 55–64 years; and ≥65 years). Assuming the prevalence of multimorbidity and medication adherence was both 50% which would give the largest sample size, and an α value of 0.05, a total of 3800 completed surveys would give a precision level of approximately 0.01. A proportional stratified sampling method was adopted according to each stratum. A total of 5114 participants were invited, 3906 residents responded, and 3866 completed interviews were conducted. According to the above age groups, the number of study participants invited in each age stratum was 183 (4.7%), 468 (12.0%), 920 (23.6%), 954 (24.4%), 689 (17.6%) and 692 (17.7%), respectively, consisting of 1690 (43.3%) male and 2216 (56.7%) female subjects. 2.3. The survey instrument This questionnaire consists of survey items in Chinese, including participants' sociodemographic details; family history of chronic diseases; self-perceived health status presented in a 4-point Likert scale (excellent; good; fair; poor); the Twenty-item Centre for Epidemiologic Studies Depression (CESD-20) Scale [26]; and the Eight-item Morisky Medication Adherence Scale (MMAS-8) [27]. The number of chronic conditions was recorded. There has been no standard approach to measure multimorbidity as it is partly subjective and dependent on the research setting [28]. We included morbidities recommended as core for any multimorbidity measure by a systematic review [29]. In the context of this study which focused on cardiovascular diseases, multimorbidity was defined as the presence of two or more chronic medical conditions [29]. These self-reported medical conditions included hypertension, diabetes mellitus, lipid disorder, myocardial infarction, stroke, chronic obstructive pulmonary disease, and cancer. 2.4. The interviews Before the surveys, a territory-wide promotional initiative was promulgated by the government and the community health centers (which are the major primary care providers in China [30]) to the whole population. The importance of the survey study was highlighted, appealing for active participation by the residents. All the interviewers underwent standardized training on the proper survey procedures and relevant communication skills. The study subjects were invited by phone to attend the community health centers closest to their residence for interviewer-administered surveys. 2.5. Quality control Before the survey study, a panel of expert consultants including epidemiologists, public health practitioners, academic professionals and physicians was set up to propose, refine and consolidate the research plan and methodology. They participated in survey design and advised on study logistics. An on-site survey team was established in each community health center, which was assigned a central coordinator. The research team members were responsible to check the completeness and accuracy of the surveys, and
discussed on any problems encountered. A pilot study was conducted in two community health centers where researchers from other centers observed the interview processes to achieve standardized survey practice. To enhance survey quality, the coordinator from each center sampled 5% of all the surveys collected and reviewed the accuracy of data by communicating with the study participants. The information of all the surveys was entered into the computer in two separate rounds to reduce the risk of error due to data entry. 2.6. Outcome measures and covariates The primary outcome measures included the presence of multimorbidity (≥ 2 medical conditions) and medication adherence. Three MMAS-8 survey sections asked adherence with antihypertensive agents; oral hypoglycemic agents; and lipid-lowering agents, respectively, among patients with physician diagnoses of hypertension; diabetes and lipid disorders who were also taking chronic medications for the corresponding diseases. Optimal adherence was defined as a MMAS-8 score N 6, out of a total score of 8 in all the sections, following the methodology used in our previous studies [31,32]. Hence if a patient was on an antihypertensive agent and a lipid-lowering drug, both MMAS-8 sections should each score N 6 for being defined as having optimal adherence. The covariates included age, sex, marital status, educational level, occupation, household income, smoking status, alcohol drinking habit, family history of chronic diseases, self-perceived health status, and depression measured by CESD-20 [normal: b15; minor to major depression: ≥ 15]. 2.7. Statistical analyses All data entry and analysis was performed by the Statistical Package for Social Sciences (SPSS) version 20.0. The characteristics of the study participants were compared according to the number of chronic conditions, by chi-square tests for categorical variables and Student's t-tests for continuous variables. A binary logistic regression model was constructed with multimorbidity as the outcome variable and all the covariates abovementioned as independent variables. In addition, among those who were taking chronic medications, adherence was used as an outcome variable and tested against each covariate consecutively in univariate analyses. Those who were tested significant (p b 0.10) were entered into a multivariate logistic regression equation. As an additional analysis, all covariates were included in another similar binary logistic regression model to detect the robustness of the findings. Interactions of variables were evaluated in each regression analysis. We performed sensitivity analysis using a cut-off of ≥6 in the MMAS-8 score to define optimal medication adherence, and repeated identical data analyses. A p value of b0.05 was regarded as statistically significant in the final regression model.
3. Results 3.1. Participant characteristics From a total of 3866 participants, the proportion of having 0, 1 and ≥ 2 medical conditions was 62.6%, 23.8% and 13.5%, respectively (Table 1). Their average age was 50.1 years (SD 14.6), and 43.2% were male subjects. Participants in the following groups had high proportions suffering from multimorbidity: male sex (15.3% vs. 12.2%); married or cohabited (13.8% vs. 11.0%); educational level at primary or lower (22.5% vs. 12.4% and 8.2% for secondary or tertiary levels, respectively); housewives or unemployed people (15.5% vs. 10.3% for employed); smokers (16.8% vs. 12.5%); positive family history of hypertension (17.7% vs. 11.2%); self-perceived health status as fair or poor (21.0% vs. 7.3%); high CESD-20 score ≥ 15 (17.3% vs. 12.8%); and poor medication adherence (45.1% vs. 37.8%). The typology of the medical conditions was shown in Fig. 1. Among all patients, the most common chronic diseases as diagnosed by physicians included hypertension (22.5%), lipid disorders (15.8%), and diabetes (10.1%). The prevalence of these three diagnoses increased sharply as the number of medical conditions increased (Fig. 1). 3.2. Factors associated with multimorbidity The proportion of poor medication adherence and depressive symptoms (CESD-20 ≥ 15) increased with the number of medical conditions (Fig. 2). From binary logistic regression analysis, older age (adjusted odds ratio [AOR] = 1.04, 95% C.I. 1.03–1.05, p b 0.001); engagement in full-time job (AOR for “others” = 0.72, 95% C.I. 0.53–0.97, p = 0.031), smokers (AOR = 1.63, 95% C.I. 1.16–2.30, p = 0.005), positive family history of hypertension (AOR = 1.51, 95% C.I. 1.19–1.93, p = 0.001), and health status self-perceived as being fair or
M.C.S. Wong et al. / International Journal of Cardiology 177 (2014) 477–482
479
Table 1 Patient characteristics (N = 3866). No medical conditions (n = 2422)b
1 medical condition (n = 921)b
≥2 medical conditions (n = 523)b
p value
50.1 (14.6)
45.4 (13.8)
56.6 (12.9)
60.3 (11.2)
b0.001
1672 (43.2) 2194 (56.8)
1001 (59.9) 1421 (64.8)
415 (24.8) 506 (23.1)
256 (15.3) 267 (12.2)
0.003
399 (10.3) 3463 (89.7)
282 (70.7) 2136 (61.7)
73 (18.3) 848 (24.5)
44 (11.0) 479 (13.8)
0.002
711 (18.4) 2495 (64.5) 659 (17.0)
341 (48.0) 1594 (63.9) 487 (73.9)
210 (29.5) 592 (23.7) 118 (17.9)
160 (22.5) 309 (12.4) 54 (8.2)
b0.001
1279 (33.1) 562 (14.6) 2020 (52.3)
899 (70.3) 357 (63.5) 1163 (57.6)
248 (19.4) 118 (21.0) 553 (27.4)
132 (10.3) 87 (15.5) 304 (15.0)
b0.001
762 (19.7) 1730 (44.8) 963 (24.9) 406 (10.5)
482 (63.3) 1053 (60.9) 607 (63.0) 277 (68.2)
170 (22.3) 426 (24.6) 232 (24.1) 91 (22.4)
110 (14.4) 251 (14.5) 124 (12.9) 38 (9.4)
0.068
2917 (75.5) 949 (24.5)
1856 (63.6) 566 (59.6)
697 (23.9) 224 (23.6)
364 (12.5) 159 (16.8)
0.003
2678 (69.3) 1187 (30.7)
1667 (62.2) 755 (63.6)
638 (23.8) 282 (23.8)
373 (13.9) 150 (12.6)
0.533
1399 (36.2) 2467 (63.8)
780 (55.8) 1642 (66.6)
372 (26.6) 549 (22.3)
247 (17.7) 276 (11.2)
b0.001
535 (13.8) 3329 (86.2)
318 (59.4) 2102 (63.1)
128 (23.9) 793 (23.8)
89 (16.6) 434 (13.0)
0.067
2116 (54.7) 1749 (45.3)
1526 (72.1) 896 (51.2)
435 (20.6) 485 (27.7)
155 (7.3) 368 (21.0)
b0.001
3252 (84.3) 606 (15.7)
2068 (63.6) 349 (57.6)
767 (23.6) 152 (25.1)
417 (12.8) 105 (17.3)
0.004
712 (18.4) 357 (9.2) 2797 (72.3)
N/A N/A 2422 (86.5)
391 (54.9) 222 (62.2) 308 (11.0)
321 (45.1) 135 (37.8) 70 (2.5)
b0.001
All patients (N = 3866)a Age (mean ± SD) Sex Male Female Marital status5 missing Single/divorced Married/cohabited Educational level Primary or below Secondary Tertiary or above Occupation5 missing Governmental/non-governmental Housewife/unemployed Othersc Household income/year (RMB$)5 missing b10,000 10,000–29,999 ≥30,000 Refuse to answer Smoking Non-smoker Smoker Drinking1 missing Non-drinker Drinker Family history Hypertension No hypertension Family history Diabetes No diabetes Self perceived health status Excellent or good Fair or poor Depression (CESD-20)8 missing Normal (b15) Minor to major (≥15) Adherence with medications Poor (MMAS-8 ≤ 6) Optimal (MMAS-8 N 6) Not on medications
CESD-20: Twenty-item Centre for Epidemiological Studies Depression Scale; MMAS-8: Eight-item Morisky Medication Adherence Scale. a The proportions across columns. b The proportions across rows. c Others include students, retired, farmers, and self-employed subjects.
Fig. 1. The typology of medical conditions.
480
M.C.S. Wong et al. / International Journal of Cardiology 177 (2014) 477–482
poor (AOR = 2.15, 95% C.I. 1.69–2.74, p b 0.001) were associated with multimorbidity (Table 2). The Nagelkerke R2 was 0.458, implying that the covariates explained for 45.8% of the outcome variability. 3.3. Factors associated with poor medication adherence From univariate analysis, younger age, engagement in full-time job, lower household income, alcohol drinking, positive family history of hypertension or diabetes, health status self-perceived as fair or poor, and the number of medical conditions ≥ 2 were associated with poor medication adherence (Table 3). In multivariate analysis, younger age (AOR = 0.98, 95% C.I. 0.96–0.99, p b 0.001); lower household income (AOR for income b RMB$ 10,000 = 0.55 to 0.57); alcohol drinking (AOR = 1.92, 95% C.I. 1.33–2.79, p = 0.001); fair or poor selfperceived health status (AOR = 1.51, 95% C.I. 1.15–1.99, p = 0.003); and the presence of two or more medical conditions (AOR = 1.34, 95% C.I. 1.02–1.77, p = 0.037) were independently associated with poor adherence with medications. There were no variable interactions among all the covariates, and all regression analyses did not show multicollinearity. Sensitivity analysis using a cut-off of ≥6 in the MMAS-8 score did not change the conclusions drawn.
develop hypertension-related chronic diseases [35]. The observed association between multimorbidity with self-perceived health status is clinically plausible, as those with multiple medical conditions are more likely to self-rate their health conditions as poor. Turning to the factors associated with poor medication adherence, a previous cross-sectional survey conducted in the primary care setting of Hong Kong reported that younger age, shorter duration of antihypertensive agents used, and poor or very poor self-perceived health status were explanatory variables for poor adherence [20]. The association between younger age and poorer drug adherence was also echoed by previous studies using database evaluations [21,23–25]. The negative association between household income and poor medication adherence was, again, compatible with a previous large-scale territory-wide study using patient health records [25]. Therefore, the associated factors identified in this survey were largely consistent with findings from existing studies.
Table 2 Factors associated with multimorbidity (N = 3866; R2 = 0.458).
4. Discussion 4.1. Statement of the principal findings From this survey involving 3866 residents of the general public, it was found that multimorbidity was associated with poor medication adherence. Older age; smoking; family history of hypertension; and fair or poor self-perceived health status were associated with multimorbidity. Younger age, low household income, alcohol drinking, and self perception of fair or poor health status were associated with poor medication adherence. 4.2. Relationship with literature and explanation of findings Our previous evaluations on the topic of multimorbidity used clinical data from electronic database [17], and surveys among the general resident population [18] and adult patients seen in primary care [19]. The prevalence of multimorbidity increased with age, and this is compatible with findings from other studies [1,33]. Other associated factors of multimorbidity identified in this study, namely smoking and poorer self-perceived health status, were also consistent with reports from published literature [34]. It has been well recognized that smokers and those with family history of hypertension are more susceptible to
Fig. 2. The association between the number of medical conditions, medication adherence and depression.
Age (mean ± SD) Sex Male Female Marital status5 missing Single/divorced Married/cohabited Educational level Primary or below Secondary Tertiary or above Occupation Governmental/non-governmental Housewife/unemployed Othersa Household income/year (RMB$) b10,000 10,000–29,999 ≥30,000 Refuse to answer Smoking Non-smoker Smoker Drinking Non-drinker Drinker Family history of hypertension No Yes Family history of diabetes No Yes Self perceived health status Excellent or good Fair or poor Depression (CESD-20)8 missing Normal (b15) Minor to major (≥15) Adherence with medications Optimal (MMAS-8 N 6) Poor (MMAS-8 ≤ 6) Not on medications Adherence with medicationsb Optimal (MMAS-8 ≥ 6) Poor (MMAS-8 b 6) Not on medications
Adjusted odds ratio
95% C.I.
p value
1.04
1.03–1.05
b0.001
1.00 (reference) 0.80
0.57–1.11
0.173
1.00 (reference) 1.26
0.82–1.93
0.299
1.00 (reference) 0.82 0.75
0.61–1.10 0.47–1.20
0.344 0.176 0.225
1.00 (reference) 1.17 0.72
0.77–1.79 0.53–0.97
0.005 0.454 0.031
1.00 (reference) 1.02 1.01 0.74
0.75–1.39 0.70–1.46 0.46–1.18
0.530 0.904 0.964 0.206
1.00 (reference) 1.63
1.16–2.30
0.005
1.00 (reference) 0.84
0.61–1.15
0.271
1.00 (reference) 1.51
1.19–1.93
0.001
1.00 (reference) 1.26
0.91–1.75
0.157
1.00 (reference) 2.15
1.69–2.74
b0.001
1.00 (reference) 1.04
0.77–1.39
0.821
1.00 (reference) 1.35 0.06
1.02–1.78 0.04–0.09
b0.001 0.037 b0.001
1.00 (reference) 1.39 0.07
1.07–1.81 0.05–0.09
b0.001 0.014 b0.001
CESD-20: Twenty-item Centre for Epidemiological Studies Depression Scale; MMAS-8: Eight-item Morisky Medication Adherence Scale. a Others include students, retired, farmers, and self-employed subjects. b Sensitivity analyses where the cut-off value of medication adherence was adjusted.
M.C.S. Wong et al. / International Journal of Cardiology 177 (2014) 477–482
481
Table 3 Factors associated with poor medication adherence (MMAS b 6; N = 1073a). Univariate analysis
Age (mean ± SD) Sex Male Female Marital status5 missing Single/divorced Married/cohabited Educational level Primary or below Secondary Tertiary or above Occupation Governmental/non-governmental Housewife/unemployed Othersb Household income/year (RMB$) b10,000 10,000–29,999 ≥30,000 Refuse to answer Smoking Non-smoker Smoker Drinking Non-drinker Drinker Family history of hypertension No Yes Family history of diabetes No Yes Self perceived health status Excellent or good Fair or poor Depression (CESD-20)8 missing Normal (b15) Minor to major (≥15) No. of medical conditions 0 or 1 ≥2
Multivariate regression analysis including significant covariates
Multivariate regression analysis including all covariates
Crude odds ratio
95% C.I.
p value
Adjusted odds ratio
95% C.I.
p value
Adjusted odds ratio
95% C.I.
p value
0.97
0.96–0.98
b0.001
0.98
0.96–0.99
b0.001
0.98
0.96–0.99
b0.001
1.00 (reference) 1.10
0.85–1.41
0.477
1.00 (reference) 1.35
0.98–1.86
0.067
1.00 (reference) 1.26
0.87–1.84
0.221
1.00 (reference) 1.04
0.66–1.63
0.872
1.00 (reference) –
–
–
1.00 (reference) 0.97
0.59–1.60
0.915
1.00 (reference) 1.20 0.97
0.90–1.61 0.62–1.52
0.331 0.205 0.877
1.00 (reference) – –
– –
– –
1.00 (reference) 1.04 0.69
0.74–1.45 0.40–1.18
0.218 0.844 0.173
1.00 (reference) 0.93 0.61
0.59–1.45 0.44–0.85
0.003 0.737 0.003
1.00 (reference) 0.98 0.78
0.59–1.63 0.54–1.12
0.268 0.942 0.178
1.00 (reference) 0.93 0.74
0.55–1.57 0.51–1.07
0.193 0.773 0.110
1.00 (reference) 0.60 0.67 0.79
0.42–0.86 0.45–1.01 0.46–1.37
0.040 0.005 0.054 0.401
1.00 (reference) 0.55 0.57 0.91
0.37–0.80 0.37–0.88 0.52–1.61
0.006 0.002 0.010 0.755
1.00 (reference) 0.55 0.60 0.92
0.37–0.80 0.38–0.94 0.52–1.63
0.009 0.002 0.026 0.783
1.00 (reference) 1.17
0.87–1.57
0.304
–
1.00 (reference) 0.91
0.61–1.37
0.654
1.00 (reference) 1.64
1.22–2.21
0.001
1.00 (reference) 1.92
1.33–2.79
0.001
1.00 (reference) 1.99
1.36–2.93
b0.001
1.00 (reference) 1.36
1.05–1.76
0.020
1.00 (reference) 1.11
0.84–1.47
0.460
1.00 (reference) 1.11
0.84–1.47
0.459
1.00 (reference) 1.57
1.07–2.29
0.020
1.00 (reference) 1.14
0.76–1.70
0.541
1.00 (reference) 1.14
0.76–1.72
0.517
1.00 (reference) 1.55
1.19–2.00
0.001
1.00 (reference) 1.51
1.15–1.99
0.003
1.00 (reference) 1.52
1.15–2.01
0.004
1.00 (reference) 1.12
0.80–1.56
0.502
1.00 (reference) –
–
–
1.00 (reference) 1.07
0.75–1.52
0.709
1.00 (reference) 1.34
1.03–1.73
0.028
1.00 (reference) 1.34
1.02–1.77
0.037
1.00 (reference) 1.34
1.01–1.77
0.042
1.0 (reference) –
CESD-20: Twenty-item Centre for Epidemiological Studies Depression Scale; MMAS-8: Eight-item Morisky Medication Adherence Scale. a There were 1078 patients taking chronic medications with 5 patients missing demographic variables. b Others include students, retired, farmers, and self-employed subjects.
The association between poor medication adherence and multimorbidity is, nevertheless, a novel finding in this study. The underlying reason is thus far unknown, and the exact relationships will need to be further explored. Patients with poor adherence could develop more complications due to poorer disease control. Alternatively, those with multiple medical conditions might encounter difficulties to comply with their daily medication-taking schedule, probably due to poorer bodily function and the higher possibility of polypharmacy, referrals or adverse drug events [36]. An additional explanation is that whilst evidence-based models in clinical consultations work well for single diseases, they could lead to “siloing” of care for patients with multiple medical conditions — leading to chaotic and fragmented care [37].
this is a cross-sectional survey and one could not draw conclusions on cause-and-effect relationships. Secondly, we have used patients' selfreport of their medical conditions which might lead to recall biases — although self-reporting was commonly used in existing literature on multimorbidity. In addition, the response rate of this population survey was only 76.4% (3906/5114), and replies on the MMAS-8 survey items might suffer from some inaccuracies with the possibility of social desirability bias. Finally, we have mainly included cardiovascular comorbidities in this study, as this survey focused on adherence with cardiovascular medications. Future evaluations should explore the relationship between drug adherence and comorbidities involving other organ systems. 4.4. Implications to clinical practice and health policy
4.3. Study limitations This is a population-based survey involving a large number of residents recruited by a representative sampling design, which enhances its generalizability to the general public. The use of validated tools to measure medication adherence and the adoption of standard methodology to assess multimorbidity were some of the study strengths. Several limitations should however be addressed. Firstly,
The clinical implications of this study are clear. From a physician's point of view, the monitoring of medication adherence should be more regular and meticulous among patients with multiple medical conditions — as were those with risk factors for poor adherence identified in this study. The effectiveness of interventions aimed at improving outcomes in patients with multimorbidity in primary care was recently reviewed [38], and it was recommended that these should target at
482
M.C.S. Wong et al. / International Journal of Cardiology 177 (2014) 477–482
common risk factors or specific functional difficulties. These study findings implied that reviews of medication adherence among these multimorbid patients are warranted. Some of the well-recognized strategies to enhance medication adherence include interventions to strengthen communication channels, like the use of written materials, adoption of electronic systems, telecommunication systems, phone calls and face-to-face counseling [39]. As primary care consultations for patients with multimorbidity need to be extended [40], future studies should evaluate how to build drug adherence-enhancing interventions in the longer consultations required among multimorbid patients. Acknowledgments We express our gratitude to all the participants in this study. We acknowledge the funding support from the Health Bureau of Weidong District, Pingdingshan, Henan Province, China. References [1] Barnett K, Mercer SW, Norbury M, Watt G, Wyke S, Guthrie B. Epidemiology of multimorbidity and implications for health care, research, and medical education: a cross-sectional study. Lancet 2012;380:37–43. [2] Grumbach K. Chronic illness, comorbidities, and the need for medical generalism. Ann Fam Med 2003;1:4–7. [3] Starfield B, Lemke KW, Bernhardt T, Foldes SS, Forrest CB, Weiner JP. Comorbidity: implications for the importance of primary care in “case” management. Ann Fam Med 2003;1:8–14. [4] Bayliss EA, Ellis JL, Steiner JF. Barriers to self-management and quality-of-life outcomes in seniors with multimorbidities. Ann Fam Med 2007;5:395–402. [5] Schoenberg NE, Kim H, Edwards W, Fleming ST. Burden of common multiplemorbidity constellations on out-of-pocket medical expenditures among older adults. Gerontologist 2007;47:423–37. [6] Wolff JL, Starfield B, Anderson G. Prevalence, expenditures, and complications of multiple chronic conditions in the elderly. Arch Intern Med 2002;162: 2269–76. [7] Tooth L, Hockey R, Byles J, Dobson A. Weighted multimorbidity indexes predicted mortality, health service use, and health-related quality of life in older women. J Clin Epidemiol 2008;61:151–9. [8] Wong MCS, Wang HHX, Leung MCM, Tsang CSH, Lo SV, Griffiths SM. The rising prevalence of self-reported hypertension among Chinese subjects: a population-based study from 121 895 household interviews. QJM-An Int J Med 2014 [in press]. [9] Lloyd-Jones DM, Adams R, Carnethon M, De Simone G, Ferguson TB, Flegal K. Heart disease and stroke statistics— 2009 update: a report from the American Heart Association statistics committee and stroke statistics subcommittee. Circulation 2009;119:e21-181. [10] Kearney PM, Whelton M, Reynolds K, Muntner P, Whelton PK, He J. Global burden of hypertension: analysis of worldwide data. Lancet 2005;365:217–23. [11] Lawes CM, Vander Hoorn S, Rodgers A, Hypertens IS. Global burden of bloodpressure-related disease, 2001. Lancet 2008;371:1513–8. [12] Mazzaglia G, Ambrosioni E, Alacqua M, et al. Adherence to antihypertensive medications and cardiovascular morbidity among newly diagnosed hypertensive patients. Circulation 2009;120:1598–605. [13] Valderas JM, Starfield B, Roland M. Multimorbidity's many challenges — a research priority in the UK. Br Med J 2007;334:1128. [14] Fortin M, Lapointe L, Hudon C, Vanasse A. Multimorbidity is common to family practice: is it commonly researched? Can Fam Physician 2005;51:244–5. [15] Tunstall-Pedoe H. Preventing chronic diseases. A vital investment: WHO global report. Geneva: World Health Organization92 4 1563001; 2005 200 [CHF 30.00]. [16] Stange KC. In this issue: challenges of managing multimorbidity. Ann Fam Med 2012;10:2–3.
[17] Wong MCS, Wang HHX, Cheung CSK, et al. Factors associated with multimorbidity and its link with poor blood pressure control among 223,286 hypertensive patients. Int J Cardiol 2014;177:202–8. [18] Wang HHX, Wang JJ, Wong SYS, et al. Epidemiology of multimorbidity in China and implications for the healthcare system: cross-sectional survey among 162,464 community household residents in southern China. BMC Med 2014 [in press]. [19] Wang HHX, Wong MCS, Wong SYS, et al. The profile of cardiovascular multimorbidity among patients in primary care in southern China: a cross-sectional study. Int J Cardiol 2013;163(2):S28. http://dx.doi.org/10.1016/S0167-5273(13) 70603-X [Supplement 1]. [20] Lee GKY, Wang HHX, Liu KQL, Cheung Y, Morisky DE, Wong MCS. Determinants of medication adherence to antihypertensive medications among a Chinese population using Morisky Medication Adherence Scale. PLoS One 2013;8:e62775. [21] Wong MCS, Tam WWS, Cheung CSK, et al. Medication adherence to first-line antihypertensive drug class in a large Chinese population. Int J Cardiol 2013;167: 1438–42. [22] Wong MCS, Kong AP, So WY, et al. Pharmacoepidemiological profiles of oral hypoglycemic agents among 28,773 Chinese patients with diabetes. Diabetes Res Clin Pract 2012;96:319–25. [23] Wong MCS, Jiang JY, Griffiths SM. Adherence to lipid-lowering agents among 11 042 patients in clinical practice. Int J Clin Pract 2011;65:741–8. [24] Wong MCS, Kong APS, So WY, Jiang JY, Chan JCN, Griffiths SM. Adherence to oral hypoglycemic agents in 26,782 Chinese patients: a cohort study. J Clin Pharmacol 2011;51:1474–82. [25] Wong MCS, Jiang Y, Griffiths S. Factors associated with antihypertensive drug compliance in 83,884 Chinese patients: a cohort study. J Epidemiol Community Health 2010;64:895–901. [26] Radloff LS. The CES-D scale: a self report depression scale for research in the general population. Appl Psychol Meas 1977;1:385–401. [27] Morisky DE, Ang A, Krousel-Wood M, Ward HJ. Predictive validity of a medication adherence measure in an outpatient setting. J Clin Hypertens (Greenwich) 2008; 10:348–54. [28] Fortin M, Stewart M, Poitras ME, Almirall J, Maddocks H. A systematic review of prevalence studies on multimorbidity: toward a more uniform methodology. Ann Fam Med 2012;10:142–51. [29] Diederichs C, Berger K, Bartels D. The measurement of multiple chronic diseases—a systematic review on existing multimorbidity indices. J Gerontol A Biol Sci Med Sci 2011;66:301–11. [30] Wang HHX, Wong SYS, Wong MCS, et al. Patients' experiences in different models of community health centers in southern China. Ann Fam Med 2013;11:517–26. [31] Wong MCS, Liu KQL, Wang HHX, et al. Effectiveness of a pharmacist-led drug counseling on enhancing antihypertensive adherence and blood pressure control: a randomized controlled trial. J Clin Pharmacol 2013;53:753–61. [32] Wong MCS, Wu CHM, Wang HHX, et al. Association between the eight-item Morisky Medication Adherence Scale (MMAS-8) score and glycaemic control among Chinese diabetes patients. J Clin Pharmacol 2014 [in press]. [33] Marengoni A, Angleman S, Melis R, et al. Aging with multimorbidity: a systematic review of the literature. Ageing Res Rev 2011;10:430–9. [34] Maskarinec G, Carlin L, Pagano I, et al. Lifestyle risk factors for chronic disease in a multiethnic population: an analysis of two prospective studies over a 20-year period. Ethn Dis 2007;17:597–603. [35] Mancia G, Fagard R, Narkiewicz K, et al. 2013 ESH/ESC guidelines for the management of arterial hypertension: the task force for the management of arterial hypertension of the European Society of Hypertension (ESH) and of the European Society of Cardiology (ESC). Eur Heart J 2013;34:2159–219. [36] Calderón-Larrañaga A, Poblador-Plou B, González-Rubio F, Gimeno-Feliu LA, Abad-Díez JM, Prados-Torres A. Multimorbidity, polypharmacy, referrals, and adverse drug events: are we doing things well? Br J Gen Pract 2012;62:e821–6. [37] Mangin D, Heath I, Jamoulle M. Beyond diagnosis: rising to the multimorbidity challenge. BMJ 2012;344:e3526. [38] Smith SM, Soubhi H, Fortin M, Hudon C, O'Dowd T. Managing patients with multimorbidity: systematic review of interventions in primary care and community settings. BMJ 2012;345:e5205. [39] Cutrona SL, Choudhry NK, Fischer MA, et al. Modes of delivery for interventions to improve cardiovascular medication adherence. Am J Manag Care 2010;16:929–42. [40] Kadam U. Redesigning the general practice consultation to improve care for patients with multimorbidity. BMJ 2012;345:e6202.
Contents lists available at ScienceDirect
International Journal of Cardiology journal homepage: www.elsevier.com/locate/ijcard
The association between multimorbidity and poor adherence with cardiovascular medications☆,☆☆ Martin C.S. Wong a,1, Jing Liu b,1, Shenglai Zhou c, Shiwei Li d, Xuefen Su a, Harry H.X. Wang a, Roger Y.N. Chung a, Benjamin H.K. Yip a, Samuel Y.S. Wong a, Joseph T.F. Lau a,⁎ a
School of Public Health and Primary Care, Faculty of Medicine, Chinese University of Hong Kong, Hong Kong Research Centre for Healthcare Management, School of Economics and Management, Tsinghua University, Beijing, China Beijing Anzhen Hospital, Beijing, China d Health Bureau of Weidong District, Pingdingshan, Henan Province, China b c
a r t i c l e
i n f o
Article history: Received 5 August 2014 Received in revised form 17 September 2014 Accepted 20 September 2014 Available online 28 September 2014 Keywords: Multimorbidity Medication adherence Associated factors Cardiovascular medications
a b s t r a c t Multimorbidity, defined as the presence of two or more chronic conditions, leads to a substantial public health burden. This study evaluated its association with adherence with cardiovascular medications in a Chinese population. A proportional stratified sampling was adopted to draw a representative sample of residents living in Henan Province, China. Interviewer-administered surveys were conducted by trained researchers. The outcomes included the number of chronic medical conditions, adherence with long-term medications (MMAS-8), and depressive symptoms (CESD-20). Binary logistic regression analysis was conducted to evaluate if medication adherence was associated with the presence of multimorbidity. From a total of 3866 completed surveys, the proportion of subjects having 0, 1 and ≥2 chronic conditions was 62.6%, 23.8% and 13.5%, respectively. Among 27.6% who were taking chronic medications, 66.6% had poor medication adherence (MMAS-8 score ≤ 6). From binary logistic regression analysis, subjects with poor medication adherence were significantly associated with multimorbidity (adjusted odds ratio [AOR]: 1.35, 95% C.I. 1.02–1.78, p = 0.037). Other associated factors included older age (AOR = 1.04, 95% C.I. 1.03–1.05, p b 0.001), smoking (AOR = 1.63, 95% C.I. 1.16–2.30, p = 0.005), family history of hypertension (AOR = 1.51, 95% C.I. 1.19–1.93, p = 0.001), and fair to poor self-perceived health status (AOR = 2.15, 95% C.I. 1.69–2.74, p b 0.001). Using medication adherence as the outcome variable, multimorbidity was significantly associated with poor drug adherence (AOR = 1.34, 95% C.I. 1.02–1.77, p = 0.037). Multimorbidity was associated with poorer medication adherence. This implies the need for closer monitoring of the medication taking behavior among those with multiple chronic conditions. © 2014 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Multimorbidity is defined as the coexistence of two or more chronic diseases in an individual, and has been increasingly identified as a usual phenomenon in patients [1]. A recent epidemiology study in a database of 1.75 million patients registered at 314 practices in Scotland reported that up to 23% of all people suffered from multimorbidity [1], rising sharply with age. It presents serious challenges to patients and physicians, requiring a patient-oriented approach supported by close ☆ Conflict of interest: None declared. ☆☆ All authors take responsibility for all aspects of the reliability and freedom from bias of the data presented and their discussed interpretation. ⁎ Corresponding author at: 5/F, School of Public Health and Primary Care, Prince of Wales Hospital, Shatin, New Territories, Hong Kong. Tel.: +852 2252 8727; fax: +852 2645 3098. E-mail address: [email protected] (J.T.F. Lau). 1 Co-first authors: The first two authors contributed equally to this manuscript.
http://dx.doi.org/10.1016/j.ijcard.2014.09.103 0167-5273/© 2014 Elsevier Ireland Ltd. All rights reserved.
coordination between generalists and specialists [2,3]. Patients with multimorbidity had more complicated healthcare needs [4], incur higher healthcare costs [5], utilize more health services [6], and were associated with poor clinical outcomes [7]. The recent decade has witnessed a rising incidence of cardiovascular disease (CVD) and its risk factors, including hypertension, diabetes and dyslipidemia [8–10]. They tend to cluster with one another — and arguably patients with a single disease may easily develop into multiple conditions [11]. Persistent adherence with chronic medications according to the healthcare providers' instructions is essential to achieve optimal disease control, and it has been shown that high adherence with cardiovascular medication was associated with a 38% decreased risk of CVD events when compared with lower adherence [12]. It is currently unknown whether patients with multimorbidity may be at higher risks for poor medication adherence. Recent studies mainly focused on the impact of multimorbidity on health-related outcomes
478
M.C.S. Wong et al. / International Journal of Cardiology 177 (2014) 477–482
and health service utilization. As highlighted in some recent reviews, multimorbidity has been regarded as a major research priority [13], yet the number and diversity of evaluations on multimorbidity have been insufficient to provide robust scientific evidence in guiding patient care [14]. The WHO called for action to prioritize healthcare resources for global management of chronic diseases in the next decade, and there is a universal consensus that the topic of multimorbidity should be a central focus of attention [15,16]. We have previously conducted studies on multimorbidity [17–19] and medication adherence [20–25] among hypertensive patients, yet most of them either relied on retrospective data [17,21–24]; or were small-scaled [19,20]. The primary objective of this study is to evaluate the factors associated with multimorbidity and medication adherence in a large Chinese population. Specifically, we tested the a priori hypothesis that the presence of multimorbidity was associated with poorer medication adherence. 2. Methods 2.1. Ethics statement The ethics clearance of the study was obtained from the Survey and Behavioral Research Ethics Committee of The Chinese University of Hong Kong. Informed consent was obtained prior to the study, and each participant was assured of survey confidentiality and anonymity. The study protocol conforms to the ethical guidelines of the 1975 Declaration of Helsinki. 2.2. Sampling frame and methods This study invited participants from the resident health registry of Henan Province, China. The registry consisted of 227,021 residents aged 18 years or older, representing 92.3% of the county's entire population. All the registered individuals were classified into 12 strata, including two strata according to sex (male vs. female), and six strata according to age groups (18–24 years; 25–34 years; 35–44 years; 45–54 years; 55–64 years; and ≥65 years). Assuming the prevalence of multimorbidity and medication adherence was both 50% which would give the largest sample size, and an α value of 0.05, a total of 3800 completed surveys would give a precision level of approximately 0.01. A proportional stratified sampling method was adopted according to each stratum. A total of 5114 participants were invited, 3906 residents responded, and 3866 completed interviews were conducted. According to the above age groups, the number of study participants invited in each age stratum was 183 (4.7%), 468 (12.0%), 920 (23.6%), 954 (24.4%), 689 (17.6%) and 692 (17.7%), respectively, consisting of 1690 (43.3%) male and 2216 (56.7%) female subjects. 2.3. The survey instrument This questionnaire consists of survey items in Chinese, including participants' sociodemographic details; family history of chronic diseases; self-perceived health status presented in a 4-point Likert scale (excellent; good; fair; poor); the Twenty-item Centre for Epidemiologic Studies Depression (CESD-20) Scale [26]; and the Eight-item Morisky Medication Adherence Scale (MMAS-8) [27]. The number of chronic conditions was recorded. There has been no standard approach to measure multimorbidity as it is partly subjective and dependent on the research setting [28]. We included morbidities recommended as core for any multimorbidity measure by a systematic review [29]. In the context of this study which focused on cardiovascular diseases, multimorbidity was defined as the presence of two or more chronic medical conditions [29]. These self-reported medical conditions included hypertension, diabetes mellitus, lipid disorder, myocardial infarction, stroke, chronic obstructive pulmonary disease, and cancer. 2.4. The interviews Before the surveys, a territory-wide promotional initiative was promulgated by the government and the community health centers (which are the major primary care providers in China [30]) to the whole population. The importance of the survey study was highlighted, appealing for active participation by the residents. All the interviewers underwent standardized training on the proper survey procedures and relevant communication skills. The study subjects were invited by phone to attend the community health centers closest to their residence for interviewer-administered surveys. 2.5. Quality control Before the survey study, a panel of expert consultants including epidemiologists, public health practitioners, academic professionals and physicians was set up to propose, refine and consolidate the research plan and methodology. They participated in survey design and advised on study logistics. An on-site survey team was established in each community health center, which was assigned a central coordinator. The research team members were responsible to check the completeness and accuracy of the surveys, and
discussed on any problems encountered. A pilot study was conducted in two community health centers where researchers from other centers observed the interview processes to achieve standardized survey practice. To enhance survey quality, the coordinator from each center sampled 5% of all the surveys collected and reviewed the accuracy of data by communicating with the study participants. The information of all the surveys was entered into the computer in two separate rounds to reduce the risk of error due to data entry. 2.6. Outcome measures and covariates The primary outcome measures included the presence of multimorbidity (≥ 2 medical conditions) and medication adherence. Three MMAS-8 survey sections asked adherence with antihypertensive agents; oral hypoglycemic agents; and lipid-lowering agents, respectively, among patients with physician diagnoses of hypertension; diabetes and lipid disorders who were also taking chronic medications for the corresponding diseases. Optimal adherence was defined as a MMAS-8 score N 6, out of a total score of 8 in all the sections, following the methodology used in our previous studies [31,32]. Hence if a patient was on an antihypertensive agent and a lipid-lowering drug, both MMAS-8 sections should each score N 6 for being defined as having optimal adherence. The covariates included age, sex, marital status, educational level, occupation, household income, smoking status, alcohol drinking habit, family history of chronic diseases, self-perceived health status, and depression measured by CESD-20 [normal: b15; minor to major depression: ≥ 15]. 2.7. Statistical analyses All data entry and analysis was performed by the Statistical Package for Social Sciences (SPSS) version 20.0. The characteristics of the study participants were compared according to the number of chronic conditions, by chi-square tests for categorical variables and Student's t-tests for continuous variables. A binary logistic regression model was constructed with multimorbidity as the outcome variable and all the covariates abovementioned as independent variables. In addition, among those who were taking chronic medications, adherence was used as an outcome variable and tested against each covariate consecutively in univariate analyses. Those who were tested significant (p b 0.10) were entered into a multivariate logistic regression equation. As an additional analysis, all covariates were included in another similar binary logistic regression model to detect the robustness of the findings. Interactions of variables were evaluated in each regression analysis. We performed sensitivity analysis using a cut-off of ≥6 in the MMAS-8 score to define optimal medication adherence, and repeated identical data analyses. A p value of b0.05 was regarded as statistically significant in the final regression model.
3. Results 3.1. Participant characteristics From a total of 3866 participants, the proportion of having 0, 1 and ≥ 2 medical conditions was 62.6%, 23.8% and 13.5%, respectively (Table 1). Their average age was 50.1 years (SD 14.6), and 43.2% were male subjects. Participants in the following groups had high proportions suffering from multimorbidity: male sex (15.3% vs. 12.2%); married or cohabited (13.8% vs. 11.0%); educational level at primary or lower (22.5% vs. 12.4% and 8.2% for secondary or tertiary levels, respectively); housewives or unemployed people (15.5% vs. 10.3% for employed); smokers (16.8% vs. 12.5%); positive family history of hypertension (17.7% vs. 11.2%); self-perceived health status as fair or poor (21.0% vs. 7.3%); high CESD-20 score ≥ 15 (17.3% vs. 12.8%); and poor medication adherence (45.1% vs. 37.8%). The typology of the medical conditions was shown in Fig. 1. Among all patients, the most common chronic diseases as diagnosed by physicians included hypertension (22.5%), lipid disorders (15.8%), and diabetes (10.1%). The prevalence of these three diagnoses increased sharply as the number of medical conditions increased (Fig. 1). 3.2. Factors associated with multimorbidity The proportion of poor medication adherence and depressive symptoms (CESD-20 ≥ 15) increased with the number of medical conditions (Fig. 2). From binary logistic regression analysis, older age (adjusted odds ratio [AOR] = 1.04, 95% C.I. 1.03–1.05, p b 0.001); engagement in full-time job (AOR for “others” = 0.72, 95% C.I. 0.53–0.97, p = 0.031), smokers (AOR = 1.63, 95% C.I. 1.16–2.30, p = 0.005), positive family history of hypertension (AOR = 1.51, 95% C.I. 1.19–1.93, p = 0.001), and health status self-perceived as being fair or
M.C.S. Wong et al. / International Journal of Cardiology 177 (2014) 477–482
479
Table 1 Patient characteristics (N = 3866). No medical conditions (n = 2422)b
1 medical condition (n = 921)b
≥2 medical conditions (n = 523)b
p value
50.1 (14.6)
45.4 (13.8)
56.6 (12.9)
60.3 (11.2)
b0.001
1672 (43.2) 2194 (56.8)
1001 (59.9) 1421 (64.8)
415 (24.8) 506 (23.1)
256 (15.3) 267 (12.2)
0.003
399 (10.3) 3463 (89.7)
282 (70.7) 2136 (61.7)
73 (18.3) 848 (24.5)
44 (11.0) 479 (13.8)
0.002
711 (18.4) 2495 (64.5) 659 (17.0)
341 (48.0) 1594 (63.9) 487 (73.9)
210 (29.5) 592 (23.7) 118 (17.9)
160 (22.5) 309 (12.4) 54 (8.2)
b0.001
1279 (33.1) 562 (14.6) 2020 (52.3)
899 (70.3) 357 (63.5) 1163 (57.6)
248 (19.4) 118 (21.0) 553 (27.4)
132 (10.3) 87 (15.5) 304 (15.0)
b0.001
762 (19.7) 1730 (44.8) 963 (24.9) 406 (10.5)
482 (63.3) 1053 (60.9) 607 (63.0) 277 (68.2)
170 (22.3) 426 (24.6) 232 (24.1) 91 (22.4)
110 (14.4) 251 (14.5) 124 (12.9) 38 (9.4)
0.068
2917 (75.5) 949 (24.5)
1856 (63.6) 566 (59.6)
697 (23.9) 224 (23.6)
364 (12.5) 159 (16.8)
0.003
2678 (69.3) 1187 (30.7)
1667 (62.2) 755 (63.6)
638 (23.8) 282 (23.8)
373 (13.9) 150 (12.6)
0.533
1399 (36.2) 2467 (63.8)
780 (55.8) 1642 (66.6)
372 (26.6) 549 (22.3)
247 (17.7) 276 (11.2)
b0.001
535 (13.8) 3329 (86.2)
318 (59.4) 2102 (63.1)
128 (23.9) 793 (23.8)
89 (16.6) 434 (13.0)
0.067
2116 (54.7) 1749 (45.3)
1526 (72.1) 896 (51.2)
435 (20.6) 485 (27.7)
155 (7.3) 368 (21.0)
b0.001
3252 (84.3) 606 (15.7)
2068 (63.6) 349 (57.6)
767 (23.6) 152 (25.1)
417 (12.8) 105 (17.3)
0.004
712 (18.4) 357 (9.2) 2797 (72.3)
N/A N/A 2422 (86.5)
391 (54.9) 222 (62.2) 308 (11.0)
321 (45.1) 135 (37.8) 70 (2.5)
b0.001
All patients (N = 3866)a Age (mean ± SD) Sex Male Female Marital status5 missing Single/divorced Married/cohabited Educational level Primary or below Secondary Tertiary or above Occupation5 missing Governmental/non-governmental Housewife/unemployed Othersc Household income/year (RMB$)5 missing b10,000 10,000–29,999 ≥30,000 Refuse to answer Smoking Non-smoker Smoker Drinking1 missing Non-drinker Drinker Family history Hypertension No hypertension Family history Diabetes No diabetes Self perceived health status Excellent or good Fair or poor Depression (CESD-20)8 missing Normal (b15) Minor to major (≥15) Adherence with medications Poor (MMAS-8 ≤ 6) Optimal (MMAS-8 N 6) Not on medications
CESD-20: Twenty-item Centre for Epidemiological Studies Depression Scale; MMAS-8: Eight-item Morisky Medication Adherence Scale. a The proportions across columns. b The proportions across rows. c Others include students, retired, farmers, and self-employed subjects.
Fig. 1. The typology of medical conditions.
480
M.C.S. Wong et al. / International Journal of Cardiology 177 (2014) 477–482
poor (AOR = 2.15, 95% C.I. 1.69–2.74, p b 0.001) were associated with multimorbidity (Table 2). The Nagelkerke R2 was 0.458, implying that the covariates explained for 45.8% of the outcome variability. 3.3. Factors associated with poor medication adherence From univariate analysis, younger age, engagement in full-time job, lower household income, alcohol drinking, positive family history of hypertension or diabetes, health status self-perceived as fair or poor, and the number of medical conditions ≥ 2 were associated with poor medication adherence (Table 3). In multivariate analysis, younger age (AOR = 0.98, 95% C.I. 0.96–0.99, p b 0.001); lower household income (AOR for income b RMB$ 10,000 = 0.55 to 0.57); alcohol drinking (AOR = 1.92, 95% C.I. 1.33–2.79, p = 0.001); fair or poor selfperceived health status (AOR = 1.51, 95% C.I. 1.15–1.99, p = 0.003); and the presence of two or more medical conditions (AOR = 1.34, 95% C.I. 1.02–1.77, p = 0.037) were independently associated with poor adherence with medications. There were no variable interactions among all the covariates, and all regression analyses did not show multicollinearity. Sensitivity analysis using a cut-off of ≥6 in the MMAS-8 score did not change the conclusions drawn.
develop hypertension-related chronic diseases [35]. The observed association between multimorbidity with self-perceived health status is clinically plausible, as those with multiple medical conditions are more likely to self-rate their health conditions as poor. Turning to the factors associated with poor medication adherence, a previous cross-sectional survey conducted in the primary care setting of Hong Kong reported that younger age, shorter duration of antihypertensive agents used, and poor or very poor self-perceived health status were explanatory variables for poor adherence [20]. The association between younger age and poorer drug adherence was also echoed by previous studies using database evaluations [21,23–25]. The negative association between household income and poor medication adherence was, again, compatible with a previous large-scale territory-wide study using patient health records [25]. Therefore, the associated factors identified in this survey were largely consistent with findings from existing studies.
Table 2 Factors associated with multimorbidity (N = 3866; R2 = 0.458).
4. Discussion 4.1. Statement of the principal findings From this survey involving 3866 residents of the general public, it was found that multimorbidity was associated with poor medication adherence. Older age; smoking; family history of hypertension; and fair or poor self-perceived health status were associated with multimorbidity. Younger age, low household income, alcohol drinking, and self perception of fair or poor health status were associated with poor medication adherence. 4.2. Relationship with literature and explanation of findings Our previous evaluations on the topic of multimorbidity used clinical data from electronic database [17], and surveys among the general resident population [18] and adult patients seen in primary care [19]. The prevalence of multimorbidity increased with age, and this is compatible with findings from other studies [1,33]. Other associated factors of multimorbidity identified in this study, namely smoking and poorer self-perceived health status, were also consistent with reports from published literature [34]. It has been well recognized that smokers and those with family history of hypertension are more susceptible to
Fig. 2. The association between the number of medical conditions, medication adherence and depression.
Age (mean ± SD) Sex Male Female Marital status5 missing Single/divorced Married/cohabited Educational level Primary or below Secondary Tertiary or above Occupation Governmental/non-governmental Housewife/unemployed Othersa Household income/year (RMB$) b10,000 10,000–29,999 ≥30,000 Refuse to answer Smoking Non-smoker Smoker Drinking Non-drinker Drinker Family history of hypertension No Yes Family history of diabetes No Yes Self perceived health status Excellent or good Fair or poor Depression (CESD-20)8 missing Normal (b15) Minor to major (≥15) Adherence with medications Optimal (MMAS-8 N 6) Poor (MMAS-8 ≤ 6) Not on medications Adherence with medicationsb Optimal (MMAS-8 ≥ 6) Poor (MMAS-8 b 6) Not on medications
Adjusted odds ratio
95% C.I.
p value
1.04
1.03–1.05
b0.001
1.00 (reference) 0.80
0.57–1.11
0.173
1.00 (reference) 1.26
0.82–1.93
0.299
1.00 (reference) 0.82 0.75
0.61–1.10 0.47–1.20
0.344 0.176 0.225
1.00 (reference) 1.17 0.72
0.77–1.79 0.53–0.97
0.005 0.454 0.031
1.00 (reference) 1.02 1.01 0.74
0.75–1.39 0.70–1.46 0.46–1.18
0.530 0.904 0.964 0.206
1.00 (reference) 1.63
1.16–2.30
0.005
1.00 (reference) 0.84
0.61–1.15
0.271
1.00 (reference) 1.51
1.19–1.93
0.001
1.00 (reference) 1.26
0.91–1.75
0.157
1.00 (reference) 2.15
1.69–2.74
b0.001
1.00 (reference) 1.04
0.77–1.39
0.821
1.00 (reference) 1.35 0.06
1.02–1.78 0.04–0.09
b0.001 0.037 b0.001
1.00 (reference) 1.39 0.07
1.07–1.81 0.05–0.09
b0.001 0.014 b0.001
CESD-20: Twenty-item Centre for Epidemiological Studies Depression Scale; MMAS-8: Eight-item Morisky Medication Adherence Scale. a Others include students, retired, farmers, and self-employed subjects. b Sensitivity analyses where the cut-off value of medication adherence was adjusted.
M.C.S. Wong et al. / International Journal of Cardiology 177 (2014) 477–482
481
Table 3 Factors associated with poor medication adherence (MMAS b 6; N = 1073a). Univariate analysis
Age (mean ± SD) Sex Male Female Marital status5 missing Single/divorced Married/cohabited Educational level Primary or below Secondary Tertiary or above Occupation Governmental/non-governmental Housewife/unemployed Othersb Household income/year (RMB$) b10,000 10,000–29,999 ≥30,000 Refuse to answer Smoking Non-smoker Smoker Drinking Non-drinker Drinker Family history of hypertension No Yes Family history of diabetes No Yes Self perceived health status Excellent or good Fair or poor Depression (CESD-20)8 missing Normal (b15) Minor to major (≥15) No. of medical conditions 0 or 1 ≥2
Multivariate regression analysis including significant covariates
Multivariate regression analysis including all covariates
Crude odds ratio
95% C.I.
p value
Adjusted odds ratio
95% C.I.
p value
Adjusted odds ratio
95% C.I.
p value
0.97
0.96–0.98
b0.001
0.98
0.96–0.99
b0.001
0.98
0.96–0.99
b0.001
1.00 (reference) 1.10
0.85–1.41
0.477
1.00 (reference) 1.35
0.98–1.86
0.067
1.00 (reference) 1.26
0.87–1.84
0.221
1.00 (reference) 1.04
0.66–1.63
0.872
1.00 (reference) –
–
–
1.00 (reference) 0.97
0.59–1.60
0.915
1.00 (reference) 1.20 0.97
0.90–1.61 0.62–1.52
0.331 0.205 0.877
1.00 (reference) – –
– –
– –
1.00 (reference) 1.04 0.69
0.74–1.45 0.40–1.18
0.218 0.844 0.173
1.00 (reference) 0.93 0.61
0.59–1.45 0.44–0.85
0.003 0.737 0.003
1.00 (reference) 0.98 0.78
0.59–1.63 0.54–1.12
0.268 0.942 0.178
1.00 (reference) 0.93 0.74
0.55–1.57 0.51–1.07
0.193 0.773 0.110
1.00 (reference) 0.60 0.67 0.79
0.42–0.86 0.45–1.01 0.46–1.37
0.040 0.005 0.054 0.401
1.00 (reference) 0.55 0.57 0.91
0.37–0.80 0.37–0.88 0.52–1.61
0.006 0.002 0.010 0.755
1.00 (reference) 0.55 0.60 0.92
0.37–0.80 0.38–0.94 0.52–1.63
0.009 0.002 0.026 0.783
1.00 (reference) 1.17
0.87–1.57
0.304
–
1.00 (reference) 0.91
0.61–1.37
0.654
1.00 (reference) 1.64
1.22–2.21
0.001
1.00 (reference) 1.92
1.33–2.79
0.001
1.00 (reference) 1.99
1.36–2.93
b0.001
1.00 (reference) 1.36
1.05–1.76
0.020
1.00 (reference) 1.11
0.84–1.47
0.460
1.00 (reference) 1.11
0.84–1.47
0.459
1.00 (reference) 1.57
1.07–2.29
0.020
1.00 (reference) 1.14
0.76–1.70
0.541
1.00 (reference) 1.14
0.76–1.72
0.517
1.00 (reference) 1.55
1.19–2.00
0.001
1.00 (reference) 1.51
1.15–1.99
0.003
1.00 (reference) 1.52
1.15–2.01
0.004
1.00 (reference) 1.12
0.80–1.56
0.502
1.00 (reference) –
–
–
1.00 (reference) 1.07
0.75–1.52
0.709
1.00 (reference) 1.34
1.03–1.73
0.028
1.00 (reference) 1.34
1.02–1.77
0.037
1.00 (reference) 1.34
1.01–1.77
0.042
1.0 (reference) –
CESD-20: Twenty-item Centre for Epidemiological Studies Depression Scale; MMAS-8: Eight-item Morisky Medication Adherence Scale. a There were 1078 patients taking chronic medications with 5 patients missing demographic variables. b Others include students, retired, farmers, and self-employed subjects.
The association between poor medication adherence and multimorbidity is, nevertheless, a novel finding in this study. The underlying reason is thus far unknown, and the exact relationships will need to be further explored. Patients with poor adherence could develop more complications due to poorer disease control. Alternatively, those with multiple medical conditions might encounter difficulties to comply with their daily medication-taking schedule, probably due to poorer bodily function and the higher possibility of polypharmacy, referrals or adverse drug events [36]. An additional explanation is that whilst evidence-based models in clinical consultations work well for single diseases, they could lead to “siloing” of care for patients with multiple medical conditions — leading to chaotic and fragmented care [37].
this is a cross-sectional survey and one could not draw conclusions on cause-and-effect relationships. Secondly, we have used patients' selfreport of their medical conditions which might lead to recall biases — although self-reporting was commonly used in existing literature on multimorbidity. In addition, the response rate of this population survey was only 76.4% (3906/5114), and replies on the MMAS-8 survey items might suffer from some inaccuracies with the possibility of social desirability bias. Finally, we have mainly included cardiovascular comorbidities in this study, as this survey focused on adherence with cardiovascular medications. Future evaluations should explore the relationship between drug adherence and comorbidities involving other organ systems. 4.4. Implications to clinical practice and health policy
4.3. Study limitations This is a population-based survey involving a large number of residents recruited by a representative sampling design, which enhances its generalizability to the general public. The use of validated tools to measure medication adherence and the adoption of standard methodology to assess multimorbidity were some of the study strengths. Several limitations should however be addressed. Firstly,
The clinical implications of this study are clear. From a physician's point of view, the monitoring of medication adherence should be more regular and meticulous among patients with multiple medical conditions — as were those with risk factors for poor adherence identified in this study. The effectiveness of interventions aimed at improving outcomes in patients with multimorbidity in primary care was recently reviewed [38], and it was recommended that these should target at
482
M.C.S. Wong et al. / International Journal of Cardiology 177 (2014) 477–482
common risk factors or specific functional difficulties. These study findings implied that reviews of medication adherence among these multimorbid patients are warranted. Some of the well-recognized strategies to enhance medication adherence include interventions to strengthen communication channels, like the use of written materials, adoption of electronic systems, telecommunication systems, phone calls and face-to-face counseling [39]. As primary care consultations for patients with multimorbidity need to be extended [40], future studies should evaluate how to build drug adherence-enhancing interventions in the longer consultations required among multimorbid patients. Acknowledgments We express our gratitude to all the participants in this study. We acknowledge the funding support from the Health Bureau of Weidong District, Pingdingshan, Henan Province, China. References [1] Barnett K, Mercer SW, Norbury M, Watt G, Wyke S, Guthrie B. Epidemiology of multimorbidity and implications for health care, research, and medical education: a cross-sectional study. Lancet 2012;380:37–43. [2] Grumbach K. Chronic illness, comorbidities, and the need for medical generalism. Ann Fam Med 2003;1:4–7. [3] Starfield B, Lemke KW, Bernhardt T, Foldes SS, Forrest CB, Weiner JP. Comorbidity: implications for the importance of primary care in “case” management. Ann Fam Med 2003;1:8–14. [4] Bayliss EA, Ellis JL, Steiner JF. Barriers to self-management and quality-of-life outcomes in seniors with multimorbidities. Ann Fam Med 2007;5:395–402. [5] Schoenberg NE, Kim H, Edwards W, Fleming ST. Burden of common multiplemorbidity constellations on out-of-pocket medical expenditures among older adults. Gerontologist 2007;47:423–37. [6] Wolff JL, Starfield B, Anderson G. Prevalence, expenditures, and complications of multiple chronic conditions in the elderly. Arch Intern Med 2002;162: 2269–76. [7] Tooth L, Hockey R, Byles J, Dobson A. Weighted multimorbidity indexes predicted mortality, health service use, and health-related quality of life in older women. J Clin Epidemiol 2008;61:151–9. [8] Wong MCS, Wang HHX, Leung MCM, Tsang CSH, Lo SV, Griffiths SM. The rising prevalence of self-reported hypertension among Chinese subjects: a population-based study from 121 895 household interviews. QJM-An Int J Med 2014 [in press]. [9] Lloyd-Jones DM, Adams R, Carnethon M, De Simone G, Ferguson TB, Flegal K. Heart disease and stroke statistics— 2009 update: a report from the American Heart Association statistics committee and stroke statistics subcommittee. Circulation 2009;119:e21-181. [10] Kearney PM, Whelton M, Reynolds K, Muntner P, Whelton PK, He J. Global burden of hypertension: analysis of worldwide data. Lancet 2005;365:217–23. [11] Lawes CM, Vander Hoorn S, Rodgers A, Hypertens IS. Global burden of bloodpressure-related disease, 2001. Lancet 2008;371:1513–8. [12] Mazzaglia G, Ambrosioni E, Alacqua M, et al. Adherence to antihypertensive medications and cardiovascular morbidity among newly diagnosed hypertensive patients. Circulation 2009;120:1598–605. [13] Valderas JM, Starfield B, Roland M. Multimorbidity's many challenges — a research priority in the UK. Br Med J 2007;334:1128. [14] Fortin M, Lapointe L, Hudon C, Vanasse A. Multimorbidity is common to family practice: is it commonly researched? Can Fam Physician 2005;51:244–5. [15] Tunstall-Pedoe H. Preventing chronic diseases. A vital investment: WHO global report. Geneva: World Health Organization92 4 1563001; 2005 200 [CHF 30.00]. [16] Stange KC. In this issue: challenges of managing multimorbidity. Ann Fam Med 2012;10:2–3.
[17] Wong MCS, Wang HHX, Cheung CSK, et al. Factors associated with multimorbidity and its link with poor blood pressure control among 223,286 hypertensive patients. Int J Cardiol 2014;177:202–8. [18] Wang HHX, Wang JJ, Wong SYS, et al. Epidemiology of multimorbidity in China and implications for the healthcare system: cross-sectional survey among 162,464 community household residents in southern China. BMC Med 2014 [in press]. [19] Wang HHX, Wong MCS, Wong SYS, et al. The profile of cardiovascular multimorbidity among patients in primary care in southern China: a cross-sectional study. Int J Cardiol 2013;163(2):S28. http://dx.doi.org/10.1016/S0167-5273(13) 70603-X [Supplement 1]. [20] Lee GKY, Wang HHX, Liu KQL, Cheung Y, Morisky DE, Wong MCS. Determinants of medication adherence to antihypertensive medications among a Chinese population using Morisky Medication Adherence Scale. PLoS One 2013;8:e62775. [21] Wong MCS, Tam WWS, Cheung CSK, et al. Medication adherence to first-line antihypertensive drug class in a large Chinese population. Int J Cardiol 2013;167: 1438–42. [22] Wong MCS, Kong AP, So WY, et al. Pharmacoepidemiological profiles of oral hypoglycemic agents among 28,773 Chinese patients with diabetes. Diabetes Res Clin Pract 2012;96:319–25. [23] Wong MCS, Jiang JY, Griffiths SM. Adherence to lipid-lowering agents among 11 042 patients in clinical practice. Int J Clin Pract 2011;65:741–8. [24] Wong MCS, Kong APS, So WY, Jiang JY, Chan JCN, Griffiths SM. Adherence to oral hypoglycemic agents in 26,782 Chinese patients: a cohort study. J Clin Pharmacol 2011;51:1474–82. [25] Wong MCS, Jiang Y, Griffiths S. Factors associated with antihypertensive drug compliance in 83,884 Chinese patients: a cohort study. J Epidemiol Community Health 2010;64:895–901. [26] Radloff LS. The CES-D scale: a self report depression scale for research in the general population. Appl Psychol Meas 1977;1:385–401. [27] Morisky DE, Ang A, Krousel-Wood M, Ward HJ. Predictive validity of a medication adherence measure in an outpatient setting. J Clin Hypertens (Greenwich) 2008; 10:348–54. [28] Fortin M, Stewart M, Poitras ME, Almirall J, Maddocks H. A systematic review of prevalence studies on multimorbidity: toward a more uniform methodology. Ann Fam Med 2012;10:142–51. [29] Diederichs C, Berger K, Bartels D. The measurement of multiple chronic diseases—a systematic review on existing multimorbidity indices. J Gerontol A Biol Sci Med Sci 2011;66:301–11. [30] Wang HHX, Wong SYS, Wong MCS, et al. Patients' experiences in different models of community health centers in southern China. Ann Fam Med 2013;11:517–26. [31] Wong MCS, Liu KQL, Wang HHX, et al. Effectiveness of a pharmacist-led drug counseling on enhancing antihypertensive adherence and blood pressure control: a randomized controlled trial. J Clin Pharmacol 2013;53:753–61. [32] Wong MCS, Wu CHM, Wang HHX, et al. Association between the eight-item Morisky Medication Adherence Scale (MMAS-8) score and glycaemic control among Chinese diabetes patients. J Clin Pharmacol 2014 [in press]. [33] Marengoni A, Angleman S, Melis R, et al. Aging with multimorbidity: a systematic review of the literature. Ageing Res Rev 2011;10:430–9. [34] Maskarinec G, Carlin L, Pagano I, et al. Lifestyle risk factors for chronic disease in a multiethnic population: an analysis of two prospective studies over a 20-year period. Ethn Dis 2007;17:597–603. [35] Mancia G, Fagard R, Narkiewicz K, et al. 2013 ESH/ESC guidelines for the management of arterial hypertension: the task force for the management of arterial hypertension of the European Society of Hypertension (ESH) and of the European Society of Cardiology (ESC). Eur Heart J 2013;34:2159–219. [36] Calderón-Larrañaga A, Poblador-Plou B, González-Rubio F, Gimeno-Feliu LA, Abad-Díez JM, Prados-Torres A. Multimorbidity, polypharmacy, referrals, and adverse drug events: are we doing things well? Br J Gen Pract 2012;62:e821–6. [37] Mangin D, Heath I, Jamoulle M. Beyond diagnosis: rising to the multimorbidity challenge. BMJ 2012;344:e3526. [38] Smith SM, Soubhi H, Fortin M, Hudon C, O'Dowd T. Managing patients with multimorbidity: systematic review of interventions in primary care and community settings. BMJ 2012;345:e5205. [39] Cutrona SL, Choudhry NK, Fischer MA, et al. Modes of delivery for interventions to improve cardiovascular medication adherence. Am J Manag Care 2010;16:929–42. [40] Kadam U. Redesigning the general practice consultation to improve care for patients with multimorbidity. BMJ 2012;345:e6202.
