Patient activation for self‐management among adult patients with multimorbidity in primary healthcare settings

Abstract Background and Aims Multimorbidity is a major public health and healthcare challenge around the world, including in Finland. As multimorbidity necessitates self‐management in everyday life, the effects of patient activation – a patient's knowledge, skills, and confidence in managing own health – on the capacity for self‐management warrant study, especially in primary healthcare settings. This study aimed to assess patient activation among multimorbid primary healthcare patients, identify factors associated with patient activation, and determine whether patients with low and high activation differ in terms of health and self‐management behavior, related perceptions, and health‐related quality of life (HRQoL). Methods A cross‐sectional survey was conducted among multimorbid patients who attended Finnish primary healthcare consultations (November 2019 to May 2020). The main outcome, patient activation, was assessed using the patient activation measure, PAM‐13®. Responses from 122 patients were analyzed using descriptive statistics, t‐tests, analysis of variance, linear modeling, the χ 2 test, and binary regression analysis. Results The mean score of patient activation was 56.12 (SD 12.82) on a scale 0–100 where ≤55.1 indicate low activation. The lower activation scores were significantly associated with old age, obesity, loneliness, and lower perceived health, functional ability, and vitality. Patients with low activation (47%) had significantly poorer physical activity, diets, adherence to care, and HRQoL, and significantly worse perceptions related to self‐management including motivation and energy, sense of normality, and support from physicians, nurses, and close people. Conclusion Patient activation among multimorbid outpatients was rather low. Findings indicate that patients' perceptions of their health and psychosocial factors may be important for activation and that patients with low and high activation differ with respect to several health variables. Determining patient activation in multimorbid patients may facilitate adaptation of care to better meet patient capabilities and needs in clinical settings. Knowledge of a patient's activation level may also be useful when developing interventions and care strategies for this patient group.


| INTRODUCTION
Multimorbidity, defined as the coexistence of two or more chronic conditions in the same individual, is increasingly common worldwide, and also in Finland. 1,2 The prevalence of multimorbidity varies according to the population and methods used to measure it, but approximately one-fourth of the population, and over half of the population over 65 years of age, have multimorbidity. [3][4][5][6] The risk of multimorbidity increases with age, as does the number of related conditions. 4,5 Also, an unhealthy lifestyle [7][8][9] and socioeconomic disadvantage [3][4][5]8 seem to be linked to an increased risk of developing multimorbidity. Multimorbidity is associated with many profound, and negative, outcomes such as decreased functional status, perceived health, [10][11][12] and quality of life, 11,12 as well as increased loneliness, 13,14 treatment burden, 15 and health service utilization. 10,16 Primary healthcare patients with multimorbidity represent a prominent part of the workload 11,17 more than half of the consultations 5,11 and the most of prescriptions. 5 Multimorbidity also increases the likelihood of hospital admission, length of hospital stays and readmission, and overall healthcare utilization and costs. 16 However, individuals with multimorbidity are in contact with their healthcare providers for only a fraction of their lives, while the most of time they are managing their condition on their own on a day-to-day basis. For example, they are responsible for adhering to prescribed medication regimes, self-monitoring their condition, and maintaining a healthy diet and adequate physical activity. 18,19 Selfmanagement, defined as an "individual's ability to manage the symptoms, treatment, physical and psychosocial consequences, and lifestyle changes inherent in living with a chronic condition," 20 is crucial in the care of individuals with multimorbidity. Successful self-management helps us to minimize troublesome symptoms, prevents the onset of additional illnesses, and allows patients to maximize their quality of life despite their chronic conditions. 19 However, life-long self-management of even one chronic condition can be challenging, and self-management can be complex and onerous for patients with multiple chronic conditions. This often necessitates coordination of care between different providers as well as management of complicated and demanding medical needs along with competing and potentially conflicting priorities and selfmanagement regimens. 18,21 The burden of treatment facing many multimorbid patients can thus be compared to the burden of diseases itself. 15 As a such, success of self-management is emphasized by one's ability and willingness to be involved in the care process, which is often referred to as patient activation.
Specifically, patient activation entails having the knowledge and skills to manage one's own health and healthcare, and confidence in managing health-related tasks. 22,23 According to the construct of patient activation more highly activated individuals believe that their role in managing their own health is important, have the knowledge and confidence necessary to act appropriately, and enact behaviors to maintain or improve their health. 22,23 Accordingly, previous empirical studies have found that patient activation is linked to many positive health behaviors among patients with diverse chronic conditions. More activated patients are more likely to engage in self-management behaviors including physical exercise, [24][25][26] healthy diet, 27,28 not smoking, 25,29 and following medication guidelines, and monitoring their condition. 30,31 Furthermore, studies have also found that higher activation is predictive of positive clinical outcomes including better blood glucose and pressure control 29,32 and lower body mass index (BMI), 25,33,34 and that those with higher activation tend to have better perceived health 32,35 and health-related quality of life (HRQoL). [36][37][38] Additionally, high patient activation has been shown to be associated with lower healthcare utilization 29,39-41 and costs. 42 Patient activation is also associated with varied psychological and psychosocial factors. Studies on patients with chronic conditions have shown that those with lower activation tend to experience depressive symptoms/depression 33,36,38 and anxiety 34,38 more frequently and have lower satisfaction with perceived social role 38 and social support. 33,36 Patient activation also seems to be linked to experiences of diverse positive and negative emotions and feelings related to illness 43 and managing health. 44 Hibbard and Mahoney 44 found that those with low levels of activation feel more often overwhelmed and less motivated in managing their own health than those with high levels of activation. Furthermore, some studies have shown that good perceived quality of patient-physician relationship was associated with higher patient activation. 43,45 A recent systematic scoping review confirmed that psychosocial and psychological factors seem to explain variations in patient activation, but that the role of these factors in influencing patient activation has so far little been studied. 46 Thus, previous studies in patients with a chronic condition(s) have shown the proven beneficial impact of patient activation on self-management and found several patient-related factors associated with patient activation. However, despite the need for more research on improving the care of patients with multimorbidity, a priority for global health research 47 only a few studies have investigated patient activation for self-management in this population. Thus, there are clear needs for research on the subject.
Knowledge of patient activation levels and factors associated with patient activation among multimorbid primary care patients could facilitate the identification of patients who need more support in selfmanagement and tailoring of counseling and care to meet patients' needs and capabilities, while also revealing potential risk factors for low activation. This in turn would inform interventions needed for this patient group.
Hence, the study's aim is to assess patient activation among multimorbid primary healthcare patients and identify factors related to patient activation in this population. Factors included were: sociodemographic (age, gender, education level, employment status, marital status, and living situation) and health-related and psychosocial factors (number of conditions, obesity, perceived health, perceived functional ability, perceived vitality, and perceived loneliness). In addition, the study aims to determine whether patients with low and high activation differ in terms of self-management behavior (physical activity, diet, use of alcohol and tobacco products; and adherence to care regimens) and perceptions related to selfmanagement, as well as HRQoL.
To this end, three research questions were posed: Participants were adult outpatients with multimorbidity who attended Finnish primary healthcare consultations for chronic condition management. The survey was conducted among those patients who attended consultations during the data collection period (November 2019 to May 2020) in one municipality in which all health centers participated in the study. Recruitment was performed by health personnel, mainly nurses who were responsible for the patients' chronic care, and took place in-person appointments with a nurse or a doctor. Before data collection, service managers at the participating health centers were briefed on the study and subsequently distributed information about the study to their staff.
Personnel were instructed to distribute questionnaires to all patients satisfying the eligibility criteria. The main inclusion criterion of the study was multimorbidity; the coexistence of two or more chronic conditions, all of which were either: A physical noncommunicable disease of long duration (e.g., cardiovascular disease, diabetes, asthma, or cancer), a mental health condition of long duration (e.g., a mood disorder), and an infectious disease of long duration such as HIV or hepatitis C. 47 Participants were also required to be at least 18 years old and to be sufficiently proficient in Finnish to complete a questionnaire. Questionnaires included detailed written information about the study purpose and objectives, the researchers' contact information, as well as a return postal envelope.

| Study measures and variables
The main outcome variable was patient activation, which was assessed using the patient activation measure, PAM-13 ® . PAM is widely used and validated in different patient populations. 22,23,48 PAM generates scores of 0-100, with higher scores indicating greater patient activation. Based on their scores, respondents can be divided into four developmental activation levels ranging from a passive recipient to an active manager of their own health. Both scores and levels are usable 22,23,48 (for more details, see Tables 1   and 3

| Statistical analyses
All statistical analysis was performed using IBM SPSS for Windows

Patient activation
Patient Activation Measure ® (PAM-13) ® PAM is unidimensional measure containing 13 items measuring patient activation: Self-assessed knowledge, skills, and confidence in self-management of chronic condition(s) and health. 22,23 Answering: Items are answered using 4-point Likert scale (1-4) ranging from disagree strongly to agree strongly, and an additional "not applicable" option (no score, the data is treated as missing). Scoring: The raw scores are summed (range 13-52) and then converted into continuous patient activation scale scores (using a calibration table provided by Insignia Health) between 0 and 100 where higher scores indicate greater activation. These scores can be stratified into four activation levels, with Level 1 being the least activated and Level 4 the most activated, correspond to scores of <47.1, 47.1 to 55.1, 55.2-67.0, and >67.1, respectively. These levels can also be used as cutoffs. 48 In this study Cronbach's α was 0.82

Sociodemographic information
Demographic information Age, gender, education level, employment status, marital status, and living situation (alone with partner, etc.)

Health-related and psychosocial information
Chronic conditions constituting multimorbidity Perceived functional ability A subjective rating of the respondent's experience of his/her ability to cope with meaningful and necessary daily life activities in the environment in which they live. One question. Answering: Possible answer options were good, quite good, fair, quite poor, and poor. 52

Perceived loneliness
Single question with answer options not at all, sometimes, and often Perceived sufficient number of close friends and relatives Single question with answer options sufficiently, not sufficiently, not at all Health-related quality of life (HRQoL) 15D is a generic instrument for measuring HRQoL the people's assessment of their health-related wellbeing, including 15 dimensions: mobility, vision, hearing, sleeping, eating, speech excretion, usual activities, mental function, discomfort and symptoms, depression, distress, vitality, and sexual activity. 53 Answering: Items are answered using 5-point ordinal scales. Respondent chooses from each dimension the level, which best describes her/his present health status. Scoring: Instrument combines a profile (a 15-dimensional description of persons health status) and a preference-based, single index measure. The scoring algorithm is provided by the meter administrator. 54 In this study, Cronbach's α was 0.87

Physical activity
The validity, as were variables those for which all questions were answered using the same response option (n = 6). All subsequent analyses were based solely on the non-excluded questionnaires (n = 100; see Table 2).
The sample size for this study was thus determined by availability of responses; however, a power analysis regarding PAM measurement using the final sample size of 100 at α = 0.05 showed the achieved power was 0.87 for the χ 2 test and 0.95 for analysis of variance (ANOVA; for perceived functional ability and perceived health) and thus well above the threshold of 0.80.
Patient activation was evaluated as both a continuous variable (PAM scores) and a categorical variable; the four activation levels (1-4) originally suggested were dichotomized into low activation (Levels 1 and 2) and high activation (Levels 3 and 4), in accordance with previous studies. 60,61 First, the associations of continuous PAM scores and categorial patient characteristics were assessed; More specially, the statistical significance of differences in mean PAM between groups was evaluated using independent samples t-tests for dichotomous categorical patient variables and one-way ANOVA for variables with more than two categories, with the Tukey test for post hoc comparisons. Second, differences between those with low and high activation in terms of self-management behavior and related perceptions were explored. When comparing means for low and high activation groups the independent samples t-tests were used. In addition, because PAM scores were previously found to be associated with patient-related factors, the association of patient activation level represents light aerobic activity and 5 represents high intensity. The time question is answered on a 4point scale where 1 represents less than 10 min and 4 represents >30 min. Scoring: The FIT index is calculated according to participants' responses by multiplying the scores obtained for each parameter as follows: FIT index = (points for frequency) × (points for intensity) × (points for duration). FIT scores range from 1 (low activity) to 100 (high activity), with scores of <36, 37-63, and > 64 indicating low, moderate, and high physical activity levels, respectively. 56 In this study, Cronbach's α was 0.78

Alcohol consumption
Alcohol use disorders identification test (AUDIT)-consumption is the first 3 questions of 10 question the AUDIT-instrument developed by the World Health Organization. 57 Answering: Each question has five possible answers.
Scoring: Each answer is assigned between 0 and 4 points, and the points are summed. On a scale ranging from 0-12, scores of 0 reflect no alcohol use. Scores of 4 or more in men and 3 or more in women are considered positive. Points below these numbers indicate low risk, scores of up to 5 indicate moderate risk, and scores of 6 or more indicate high risk in both genders. 58  In this study, Cronbach's α was 0.77

Self-management perceptions
26 statements on ACDI-scale 59 as follows: Energy and motivation (two items) Cronbach's α was 0.76 Sense of normality (nine items): Cronbach's α was 0.83 Fear of complications and additional diseases (two items) Cronbach's α was 0.94 Support from physicians (four items) Cronbach's α was 0.86 Support from nurses (four items) Cronbach's α was 0.89 Support from family and friends (five items) Cronbach's α was 0.80 Answering: Items are answered using 4-point Likert scales (1-4) with options totally disagree, partly disagree, partly agree, totally agree Scoring: Mean sum variables 1-4, with scores of ≥3.5 indicating good value PAUKKONEN ET AL. | 5 of 17 with self-management behaviors and perceptions was also calculated by adapting perceived health, loneliness, and obesity in a linear model. Also, because HRQoL is known to be related to age, gender, and disease count, 62 the association of activation level with HRQoL dimensions was also calculated by adjusting these factors in the linear model. Differences between proportions for categorical variables were compared using the χ 2 test. Further, binary logistic regression analysis with the calculation of odds ratios (ORs) was used to identify effects between patients' activation (low and high) and self-management behaviors and perceptions.

| Sample characteristics
The mean age of the participants was 68 years (SD 11.4). Over half (58%) of the respondents were women; 42% were men. Somewhat less than a third (29%) had only a basic education, one-fifth (19%) had completed secondary education, and a half (52%) had completed at least postsecondary education. The majority (86%) of the participants were retired. Well over half were married or in a registered partnership (61%) and half lived with a spouse or partner (52%).

| Factors associated with patient activation score
The only sociodemographic factor significantly associated with the PAM score was age, although this association appeared only when comparing  Table 4). between high-and low-activation patients ( Table 6).
HRQoL also differed significantly between the patient activation groups; more specifically, patients with low activation had a significantly worse HRQoL (p = 0.001). HRQoL dimensions with statistically significant between-group difference were breathing  Table 7).

| DISCUSSION
This study provided new knowledge about patient activation and factors associated with patient activation for self-management in patients with multimorbidity in Finnish primary healthcare settings, and differences related to self-management between patients with low and high activation.
The mean activation score for the studied population was 56.1, which was quite low. According to a recent systematic review of patient activation in people living with chronic conditions, the mean PAM score in studies ranged from 59.1 to 82.5 (including 32 articles T A B L E 3 The levels of patient activation 48 and their proportion in this study PAM level PAM score (possible range 0-100) Interpretation

Proportion in this study
"Disengaged and overwhelmed" 23% Low Level 1 <47.0 Individuals tend to be passive and lack knowledge and confidence. They may not yet understand their role in care process and managing their health.
"Becoming aware, but still struggling" 24% Level 2 ≥47.1 and ≤55.1 Individuals have some knowledge, but large gaps remain, and they still lack the confidence to manage their health. They may believe health is largely out of their control.
"Taking action" 46% High Level 3 ≥55.2 and ≤67.0 Individuals appear to be taking action and building self-management skills but may still lack the confidence and skill to maintain their behavior.
"Maintaining behaviors and pushing further" 7% Individuals have adopted many of the behaviors needed to support their health but may not be able to maintain them in the face of stress or change.
PAUKKONEN ET AL. published between 2005 and 2019). 63 However, the mean PAM score of this study is consistent with a previously reported value for multimorbid older adults. 26 Nevertheless, the proportion of participants exhibiting the highest level of activation (4) was very low (7%).
This is consistent with a study in patients with one or more selected chronic conditions that found that only a minority of patients scored PAM Level 4. 33 Similar results were also reported previously for older patients with complex medical needs 35  In this study, perceived loneliness was found to affect patient activation. More specially, feelings of loneliness and insufficient close relationships with others were more common among patients with lower patient activation. Previously, loneliness was found to be associated with low activation among military veterans with depression. 68 These results also agree with previous findings that satisfaction with social role 38 and social support 33,36 was positively associated with patient activation in patients with a chronic condition(s). However, living alone, or living situation, in general, was not associated with activation, in keeping with previous findings on multimorbid patients. 36 These findings confirm the importance of taking patients' perceptions of their social relationships into account for self-management of multimorbidity, especially when loneliness is known to be associated with multimorbidity, 13 also in physical multimorbidity alone. 14 In the studied population here, PAM was apparently unrelated to the number of chronic conditions, which is consistent with previous findings on adults with multimorbidity. 26  vitality. Lower self-rated health (based on diverse metrics) was also previously associated with low activation in multimorbid patients. 26,33,38 Another factor significantly associated with patient activation was obesity: obese participants had lower activation than their nonobese counterparts. This is consistent with previous study findings among people with type 2 diabetes mellitus in the United States 69 and reports showing that lower activation was associated with higher BMI in chronic patients. 25,33,34 Previous studies have also stated that obesity is strongly associated with multimorbidity 7,70,71 in keeping with the results obtained here: almost half of the respondents in this study fell into this category, compared to around 25% of the adult Finnish population. 51 Patients with low and high activation also exhibited significant differences in self-management behaviors: low activation participants ate significantly less frequently with a varied diet, complied less frequently with dietary instructions, and had lower frequencies of physical activity than those with high activation. Healthy eating 27 and physical activity were previously found to be associated with activation. 25,26 Diet and physical activity are key lifestyle variables in the prevention and care of chronic conditions; physical activity, in T A B L E 5 Differences in self-management behaviors between participants at different levels of patient activation (mean [SD] or n (%), p-values, and effect size) and whether activation level is an explanatory factor for behaviors (odds ratio particular, has been described as a polypill for several chronic diseases, 72 has been related to increased life expectancy, and exhibits an inverse dose-response association with mortality in the multimorbid population. 73 The results of this study also suggested that low activation in multimorbid patients predicted poorer overall adherence to care regimens. As a such, patient activation may potentially be incorporated as one tool to address the challenges of inadequate adherence, physical activity, and diets, as well as overweight and obesity, as described above.
In addition to self-management behaviors, this study examined patients' perceptions relating to self-management. High activated participants had more positive perceptions in terms of having energy and motivation to care for themselves, as well as feeling a sense of normality in care; meaning, for example, that they more often felt that self-management produces well-being and enabled them to stay healthy and was a natural part of their daily routine. Instead, lowactivation patients more frequently agreed that they did not follow recommended treatment guidelines because the guidelines did not fit their lifestyle. This is consistent with previous reports that patients with low activation may not consider their role in the care process to be important and are more likely to have low confidence in their ability to self-manage and to feel overwhelmed as a result. 74 The HRQoL analysis performed here reinforced the finding that such patients have significantly more difficulties with mental function and feel more depressed and distressed than high activation patients. In addition, patients with high activation felt more support for selfmanagement from both physicians and nurses. The relationship between activation and perceived support may be complex; perceived support may contribute to higher activation, but patients with high activation, that is, higher skills and confidence, may also find healthcare encounters more supportive. The explanation may be that they find encounters more apprehensible and moreover are more adept at getting their healthcare providers to meet their needs. 45,48 This result is supported by previous findings that a higher level of activation was associated with better patient-professional relationships experienced by patients. 43,45 In any case, finding an association between activation and perceived support was important in itself and warrants further attention.
Because a cross-sectional design was used, the findings demonstrate associations between patient activation and the studied factors but cannot be used to infer causality. However, it was speculated that the direction of the causal relationship between patient activation and health is likely to go in both directions, also suggested by Hibbard et al. 74 ; meaning that those with low patient activation are at risk for T A B L E 6 Differences in self-management perceptions at different level of patient activation (mean [SD], p-values, and effect size) and whether patient activation is explanatory factor for perceptions (odds ratio  poor self-management and health outcomes, but also that those patients who are overwhelmed by their illness or circumstances, for example, those with poor perceived health, are likely to find selfmanaging of their conditions on a day-to-day basis to be more difficult, and as a result, they score low patient activation. However, previous studies have shown that experiences of success in selfmanagement, even small, can build positive emotions and confidence and initiate positive progress. 44 Patient activation is modifiable, and its increase can also be promoted by appropriate supportive actions. Especially, interventions tailored according to patients' activation levels have previously been shown to be effective. 48 Insignia Health Inc.) or registration (15D) 54 as required. All prospective participants were given detailed written information about the purpose and objectives of the study, as well as assurances regarding anonymity, confidentiality, and the voluntary nature of participation. Furthermore, the contact details of the researchers were provided so prospective participants could ask additional questions. Completing and returning the anonymous questionnaire was considered to imply informed consent for participation in the study. The data were collected, processed, and stored without identifying information. Ethical approval was thus not required.

| CONCLUSION
This study on multimorbid primary healthcare patients showed that levels of activation, that is, the knowledge, skills, and confidence to manage one's own health and healthcare, in this population were rather low. Patient activation was negatively associated with old age, obesity, perceived loneliness and lack of close friends and relatives, and poor perceived health, functional ability, and vitality. However, it was not associated with other sociodemographic factors or the number of conditions suffered by the patient. The results presented here indicate that patients' perceptions of their health and functional ability as well as psychosocial factors may be important for activation and should be considered, rather than traditional socio-demographic factors, when assessing a patient's risk of low activation. Additionally, patients with low and high activation exhibited several differences in terms of health behaviors, perceptions related to self-management, and HRQoL. These results suggest that patient activation is important for self-management and well-being in multimorbid patients. Knowledge of a patient activation level may be useful when developing tailored support and interventions suited to their capabilities and needs, also considering individual needs to build more knowledge, confidence, and/or motivation. Moreover, the results highlight the importance of patient-centredness toward a whole person in the care of patients with multimorbidity.

| Practice implications
Patients with multimorbidity could benefit from support for patient activation to enhance self-management needed in their everyday life.
Activation support may include supporting knowledge and skills for self-management but also strengthening confidence and understand- Funding. Funding bodies had no role in the study; not in the design of the study, in the collection, analysis, and interpretation of data, in the writing of the report; or in the decision to submit the article for publication.

CONFLICT OF INTEREST
The authors declare no conflict of interest.

TRANSPARENCY STATEMENT
Leila Paukkonen affirms that this manuscript is an honest, accurate, and transparent account of the study being reported; that no important aspect of the study have been omitted; and that any discrepancies from the study planned (and, if relevant, registered) have been explained.