Abstract
The study focuses on developing a tool to measure the engagement of diabetes patients and to explore its dimensions. This study was based on descriptive research design and purposive sampling method. Personal interview was conducted among the doctors to ascertain the face and content validity of the questionnaire. Subsequently, interview was conducted with the patients and exploratory factor analysis was used to reduce the number of items in the diabetes patient engagement scale. Diabetes patient engagement measurement model resulted in seven dimensions consisting of 21 items. The dimensions patient satisfaction, patient initiatives, organized health care, informed choice, health promotion, and prevention are unidimensional factors, and the dimension self-management (routine and clinical assessment) is multidimensional. This study confirmed the validity of the new multidimensional diabetes patient engagement scale to assess and measure the intensity of engagement of the patient with their respective care towards doctor and hospital. Practical implication is to facilitate in predicting and determining the engagement level of patient with their health care provider, which increases good health outcomes and quality of life.
Introduction
Diabetes is a chronic condition that affects the body’s capacity skill to use the energy found in food. It is a progressive disorder characterized by hyperglycemia due to complete or partial deficiency of insulin hormone (World Health Organization [WHO], 1999). The WHO has declared that more than 346 million people worldwide have diabetes mellitus. This will double by 2030 without any disruption. Almost 80% of diabetes deaths occur in middle- and low-income countries. According to the WHO report, India today highlighted that there are over 3.2-crore diabetic patients. This number is projected to increase to 7.94 crores by the year 2030 (Wild et al., 2004). One of the biggest challenges faced by health care providers is to continuously engage patients with diabetes (Wagner, Austin, & Von Korff, 1996).
The great consequence on the part of the health care in the prevention of long-term complications is to make the patients realize the importance of regular follow-up of diabetes with the doctors. Besides, regular follow-up strict control on the level of blood sugar can prevent or delay in the progression of complications associated with diabetes (Ohkubo et al., 1995; Turner, 1998).
The desire of diabetic patients is to not only to control the glycemic level, but also to be consistent in the prevention of the associated problems. Some of the Indian studies revealed a very poor commitment to treatment schedule due to poor attitude toward the disease and poor health literacy among the general public (Chew, 2004; Shobana, Christina, Vijay, & Ramachandran, 2005). In the delivery of health care, making the patient a dynamic player and to engage them in the management of care reduces doctor’s efforts, the length of hospitalization, and poor outcomes (Barello, Graffigna, & Vegni, 2012; Forbat, Cayless, Knighting, Cornwell, & Kearney, 2009; Hart, 2002). The concept of patient engagement assures to deliver essential value for the patient, doctors, hospital, and other health care stakeholders.
Conceptualization
In psychology, Brodie, Juric, Ilic, & Hollebeek (2013) defines “customer engagement” as “a psychological state that occurs by virtue of interactive, cocreative customer experience with a focal agent/object,” while consumer engagement and participation are acknowledged as an upcoming field of health research and practice (Consumers Health Forum of Australia, 2010; Gregory, 2006). In a research, patient engagement was identified as “interactive, bi-directional exchange between stakeholders and researchers,” which focuses on the health questions and outcomes that are most important to patients and this improves the quality and relevancy of research (Curtis et al., 2013). Illustrious definition (Coulter, 2006) focuses on the relationship between patients and health care providers as they work together to “promote and support active patient and public involvement in health and healthcare, to strengthen their influence on healthcare decisions, at both the individual and collective levels” and the Center for Advancing Health’s (2010) engagement behavior framework focuses on behavior, defining engagement as “actions people take for their health and to benefit from health care.”
Engagement as a behavioral activation is related to healthy behaviors and positive health outcomes (Hibbard, Mahoney, Stockard, & Tusler, 2007). It is also an action that a person executes toward drug prescription and forms an essential component of high-quality health care services (Anhang Price et al., 2014) as well as sensational attachment to the health care providers (Gemmel & Verleye, 2010). Engagement is a two-sided affair. Patient engagement occurs when physician encourages, requests, and recognizes the patient’s perspective (Wilkins, 2012).
Dynamic, active engagement leads to an improved bond between consultants and patients, which facilitates the implementation of a treatment plan (Viederman, 2002). Although an informed and engaged consumer plays a most important role in enlightening the quality of care, he or she enables patient alliance with clinicians and enhances recovery experience (Simpson & Joe, 2004).
Involving people in healthcare is very powerful. Engagement is multi-dimensional. It can take the form of a very personal event, such as a shared decision between an individual patient and healthcare professional; or it can be a systematic public or family experience It can be targeted at improving provider performance, as when patient and family members and healthcare professionals redesign healthcare services together. It can begin with simple sharing of information, move on to the conversation, and progress into partnership. Whatever kind of form it takes, engagement changes the focus, from taking action to develop the health and care for the people, to taking action along with the people—a simple yet radical notion. . . Engagement is a mutual relationship. (Edgman-Levitan, 2013)
Obviously, a chronically diseased patient can manage and engage by making the right decision, following regular medication and self-monitoring his or her status through proper self-care and exercise (Chodosh, 2005), whereas perception of patient/family advisors in rapid process improvement workshops (RPIWs) on the engagement of consumers in health care, shaped and promoted new forms in expression, forms, and spaces of power (Goodridge, Isinger, & Rotter, 2018). The mutual understanding of patient engagement has the involvement and challenges on a different framework such as jurisdictions, health system, specialty, discipline, the background of patients, and time period encountered by both patients and health care professionals (Wong, Lui, Cheung, & Yam, 2017). As a result of diabetes, patient needs engagement with their health care providers constantly to lead a quality life.
Previously, patient engagement surveys address some aspects of integrated care, none do so comprehensively, and its measurement is considered an emerging field. At present, no validated, self-administered, diabetes patient engagement (DPE) questionnaire exists for engagement of diabetes patient with their health care providers. Most of the studies on patient engagement were qualitative. Only a few tools exist. Alta rum Consumer Engagement (ACE)TM helps to identify the measure of individual engagement with health and health care decision (Bleich, Ozaltin, & Murray, 2009). There are several instruments that measure one or more aspects of engagement such as patient activation (Hibbard, Mahoney, Stockard, & Tusler, 2005) and shared decision making (Simon et al., 2006). Hence, we have developed the first self-administered, DPE questionnaire and this case outlines the process of developing the DPE scale.
Conceptual framework
The conceptual model underlying the DPE was developed using a comprehensive list of items that represent the various factors of patient engagement. The scale items for DPE were selected from the studies, which were directly or indirectly related to patient engagement. For this purpose, diverse patient engagement survey reports and published articles were taken into consideration.
The following articles were selected as a part of an extensive review of literature:
Engagement behavior framework (EBF; Center for Advancing Health, 2010),
Patient activation measure (PAM; Hibbard et al., 2005),
Regional primary care coalition (RPCC; Langston, Pattabiraman, & Kaye, 2012),
Picker’s questionnaire for “what do you think about your Dr?” (Chisholm & Askham, 2006),
King’s fund—the quality of patient engagement (Parsons, Winterbottom, Cross, & Redding, 2010),
Eight dimensions of patient-centered care (Bender, 2013),
Picker’s “Engaging patient in health care” (Coulter, 2006),
NeHC stakeholder survey on patient engagement (Perna, 2012).
Three major dimensions were identified for patient engagement from these articles and survey reports. These dimensions were grouped as doctor-related factors, patient-related factors, and hospital-related factors.
The first factor, “Doctor related” comprised of six constructs, namely, Communication with the doctors (eliciting information from patients about their symptoms by listening, discussing, encouraging, and answering questions), Informed choice (reliable with the decision maker’s values, and execute behaviorally), Shared decision making (decisions that are shared by doctors and patients, updates the best facts available and provides information according to the specific characteristics and values of the patient), Participation in treatment (the involvement of the patient in the decision-making process regarding health issues), The empathy of doctors (doctors understanding the patient’s situation, feelings, and motives and emotional state of mind), Patient preference (patient-specific inputs to the care process).
The second factor, “Patient-related” comprised of six constructs, namely, Patient satisfaction (satisfaction with the service provided by the practitioners, nurses, accessibility, facilities and other technical staff, in short satisfaction with the service as a whole), Health literacy (a group of skills, such as the ability to perform basic reading, educational knowledge about the disease), Self-management (individual assume preventive or therapeutic health care activities often in collaboration with health care professionals), Promotion of health (set priorities to optimize health and prevent disease and act on them), Prevention of disease (self-interest of the person to promote, protect, and rebuilt health, when it is impaired, and to minimize suffering and distress), Organized health care (making organized appointments, arrange for assistance, bringing documentation, bringing previous medical a summary and recent test results to visits as appropriate).
The third factor, “Hospital related” comprised of constructs such as Partnering with patients (associating patients to improve health care in the community and learns how the health care system works), Engagement in community health (patients are more engaged in advocating for the health of their communities), Patient centeredness (providing care toward individual patient’s choice, requirements, and value).
Method
Instrument development
DPE survey was developed based on a conceptual framework. Potential items for the scale were identified through a review of existing patient engagement surveys for related constructs. Then, iterative discussion and refinement of the constructs were done with expert-panel members to identify the components of our theoretical framework that were not covered satisfactorily by existing surveys, adapt existing items that represented our dimensions, and create new items to fill conceptual gaps. Finally, 254 items were generated, which represented the 15 dimensions as the most essential factors in our framework.
Exploratory interview (pretest)
The second step was the collection of qualitative data by interviewing physicians, diabetologists, and dietitians practicing in various specialty hospitals in India. The aim of this interview was to ensure that the items bring out appropriate responses and uncover ambiguous wording or errors before the survey is launched at large. It was done in two stages.
In the first stage, a small set of 30 physicians, diabetologists, and some dieticians were interviewed from July 2014 to November 2014 with a listing of all 254 initial items on the 3-point scale (not representative, somewhat representative, and clearly representative).
This resulted in the reduction of the initial 254 items to 110 items in the first phase.
In the second stage, two subsequent rounds of discussion with four medical experts and eight specialized medical experts were done, which resulted in the reduction of 110 items to 72 items and then to 61 items.
These 61 items were compiled into a questionnaire and a pretest survey was conducted with 200 diabetes outpatient volunteers personally from December 2015 to March 2015. The patients were asked to look for any difficulties with wordings and to recheck on the content and face validity.
Pilot study
A pilot study was done in leading diabetes hospitals in India from August 2015 to September 2015 in India. With prior intimation and permission from the hospital authorities, patients who have already received the care from the respective hospitals up to that time from the diabetologists were purposively selected and contacted personally.
Statistical analysis
The data were analyzed using IBM SPSS Version 21. The items were analyzed using Exploratory Factor Analysis (EFA), reliability analysis, validity analysis, and Confirmatory Factor Analysis (CFA). Factor analysis was used to extract component factors with varimax rotation. Two basic methods are used for extracting factors in EFA, namely, common factor analysis and principal component factor analysis, whereas common factor analysis is for exploring the latent dimensions represented in the original variables and principal component factor analysis is used mainly for item reduction (Conway & Huffcutt, 2003).
Results
A total of 730 were found to be valid out of a total of 850. It is generally recommended that a sample size of at least 300 cases were needed for factor analysis (Tabachnick & Fidell, 2007). Factor analysis for the pretest data yielded a high Cronbach’s alpha score of above .80, which resulted in the reduction of 61 items to 54 items, which were grouped under 10 dimensions, namely, communication (16 items), patient satisfaction (six items), patient initiatives (five items), self-management (five items), health promotion (four items), prevention (five items), organized health care (three items), informed choice (three items), diet management (four items), and empathy (three items).
The 54 items were further subjected to EFA using principal axis factoring with varimax rotation, interitem correlation (less than .3 or more than .9), and anti-image (.50) as the criterion. Diet management dimension with four items was removed whose Cronbach’s alpha value was .622, which was less than .70 and one item in organized care was also less than .45 and hence removed. In addition to this, 14 more items were also removed. So in total, 19 items were eliminated.
Then the scale comprising of 35 items were grouped under nine dimensions, namely, communication (11 items), patient satisfaction (five items), patient initiatives (three items), self-management (four items), health promotion (two items), prevention of diseases (two items), organized health care (two items), informed choice (three items), and empathy (three items). The factor loading and communality values of the retained items are given in Table 1.
KMO, factor loading (FL), anti-image (AI), Eigenvalue (E), variance extracted (V), and Communality Value (CV) of diabetes patient engagement.
KMO: Kaiser–Meyer–Olkin test.
Reliability analysis
The reliability of the nine constructs was excellent and the Cronbach’s alpha value was .937, which indicates acceptable reliability coefficient (Nunnally & Bernstein, 1978). The reliability of the data was also examined for split subsamples, obtained by dividing the respondents into two equal halves, to know whether variation in both the halves is within the range of sampling. The alpha value for the subsamples ranged between .879 and .941 and supports the internal consistency of the scale (Table 2).
Cronbach’s alpha value, split half, and constructs reliability of patient engagement scale.
Validity analysis
The content and face validity of the scale was duly assessed through review of literature and deliberations with doctors, patient, and experts for the selection of an item in the questionnaire.
Construct validity
The KMO (Kaiser–Meyer–Olkin) measure of sampling adequacy value was .935, Bartlett’s test of sphericity and variance explained was 56.956%, which indicates the relevance of factor analysis for factor identification, which consequently checked the construct validity (Netemeyer, Bearden, & Sharma, 2003).
Convergent validity
The factor loading and communality value of the items in nine dimensions ranges from .45 to .73, which indicates convergent validity.
Assessment and purification of DPE scale—CFA
After purifying the scale items using EFA and reliability analysis, the final factors emerged in EFA, are then analyzed in CFA to refine them further vis-a-vis to confirm whether the structure emerged is perfectly fit or not for CFA analysis. As such measurement model is estimated using maximum likelihood estimation (AMOS, Figure 1). The patient engagement scale is finalized in three stages.
Patient engagement scale.
First stage—measurement model
In first stage measurement model is analyzed using critical ratio (CR) values above 1.96 at 95% level of significance, regression weights (RWs) above .50 and squared multiple correlations (SMCs) above .50 as criteria (Netemeyer et al., 2003). This also helped in ensuring the unidimensionality of the constructs (Table 3).
CR, SRW, and SMC values of patient engagement scale.
COMM: communication; PS: patient satisfaction; PI: patient initiatives; SM: self-management; ORG: organized health care; IC: informed choice; EMP: empathy; HP: health promotion; PREV: prevention; CR: critical ratio; SRW: standardized regression weight; SMC: squared multiple correlation.
DPE scale comprised of nine factors with two multidimensionality factors, namely, communication and self-management and seven unidimensionality factors such as patient satisfaction, patient initiatives, organized health care, informed choice, empathy, health promotion, and prevention program. All the items of these dimensions are scored on 5-point Likert-type scale (1 = strongly disagree, 5 = strongly agree).
The first multidimensional factor, communication, comprised of 11 items with three subfactors: quality of rapport (five items), informed consent (three items), patient’s compliance (three items).
The second factor, patient satisfaction, contains five significant items.
The third factor, patient initiative, contains three items.
The fourth factor, named self-management, contains two subfactors, namely, routine assessment (two items) and clinical assessment (two items).
The fifth factor, which was termed as organized health care, contains two items.
The sixth factor, informed choice, comprises of three significant items.
The seventh factor, empathy, contains three items.
The eight factor, health promotion, contains two items.
The ninth factor, prevention program, contains two items.
Second stage—reliability of the patient engagement scale
In the second stage, the nine-factor model structure for patient engagement scale is examined for the psychometric properties using confirmatory analysis. Composite reliability (above .7) and average variance extracted (AVE; above .50 standardized loading) and discriminant validity by comparing the AVE with SMCs were assessed.
Construct reliability
The construct reliability of the constructs/subscales was assessed (Table 4). The composite reliability of the entire construct was greater than .7. The result indicates that the scale is quite reliable.
Discriminant validity, average variance extracted, construct reliability, and convergent validity of patient engagement scale.
SM: self-management; Patsat: patient satisfaction; Patinit: patient initiative; ORG: organized health care; IC: informed choice; Prev: prevention; HP: health promotion; CR: critical ratio; AVE: average variance extracted.
The AVE is a measure of the amount of variance captured by a construct from each scale. The AVE with recommended values of .50 or higher provides evidence for convergent validity (Fornell & Larcker, 1981). The AVE value of the patient engagement dimensions arrived at .607 (self-management), .577 (patient satisfaction), .575 (patient initiative), .769 (organize health care), .557 (informed choice), .684 (prevention), and .548 (health promotion) indicate validity of scale (Table 4).
Nomological validity
To establish nomological validity, a three-items patient loyalty (PL), that is, “provide positive word of mouth about this hospital to other people,” “ recommend this hospital to others,” “Willing to reuse the services of hospital” were used to examine patient engagement and PL (CR = 10.27, SRW = .704, SMC = .496). The relationship was found to be significant. As such the patient engagement scale possesses nomological validity.
Discussion
The developed and validated DPE scale is a new, specific, self-administered measure for evaluating the level of engagement of diabetes patient with their doctors and health care provider. On the basis of psychometric testing, the DPE was judged to be reliable and valid as a questionnaire for engaging patients. Regarding reliability, good to excellent internal consistency and reproducibility were observed in all the subdomains.
Regarding validity, each of the factors and the subscales had adequate internal consistency, a valid construct, convergent with acceptable reliability coefficient. Regarding factor analysis, unexpected loadings showed insufficient internal consistency for diet management, organized health care and other items and, hence, could not be incorporated in the scale. Concerning the confirmatory analysis, the dimension communication and empathy does not fulfill the minimum AVE and construct reliability, and therefore, the dimensions with the respective items were also excluded. Overall, the findings suggest that self-management, patient satisfaction, patient initiative, organized health care, informed choice, prevention, and health promotion factors contribute more to the engagement of diabetes patients toward their doctors and their hospital.
Implications for research and practice
The DPE will be a useful tool for both the patient and health care providers in designing and inculcating an engagement pattern that can improve especially the quality of life of aged people. The measure can be used in a hospital setting to assess individual patient’s engagement and to develop care plans tailored to that patient and integrate that into the processes of their care. Future researches can focus on the engagement of patients with other noncommunicable, chronic diseases such as cardiac, cancer, renal, and so on.
Limitation and future research
The current study is the first to evaluate the engagement among the patients with their health care providers using a DPE scale. Increase of diabetes incidence across the countries necessitates the people to manage their committed lifestyle and most of the patients need assistance to manage their disease with their health care providers, which is a difficult task for an individual (English, Masom, & Whitman, 2018). Thus, this study helps in exploring the status of diabetes patients to engage with their physician or health care providers and adhere to the treatment. Also, this study helps in proper communication with their physician and understanding about the disease status. However, further analysis is needed to explore the strength of the evidence on a different set of patients and in other countries. However, these findings are not limited by culture-specific factors, and hence, is applicable across countries.
