Patient communication pattern scale: psychometric characteristics

Abstract Background In western societies, a shared decision‐making model for doctor–patient relationships calling for open and collaborative communication is recommended. Research focuses mainly on the doctor's communication patterns, while research on patient communication patterns is rare. The purpose of this study was to develop a tool for evaluating patient's communication patterns – the Patient Communication Pattern Scale (PCPS). Methods Interviews based on structured questionnaires were conducted with 251 cancer patients. In addition to the 14‐item PCPS, the questionnaire included questions regarding education, religiosity and desirability of control in general and over one's own health in particular, for validating the scale. Results The PCPS was found to be a valid and reliable tool for evaluating patients' communication patterns. Confirmatory factor analysis supported the PCPS designed structure of five facets: (1) Information, (2) Clarification, (3) Initiation, (4) Preferences and (5) Emotions. Conclusion The PCPS is a reliable scale for evaluating patient communication patterns. The use of this scale can assist in promoting related research and in developing interventions for enhancing open and collaborative doctor–patient communication.


Introduction
In recent decades, significant changes have occurred in approaches concerning the doctorpatient relationship. This relationship has evolved from the paternalistic model dominant in the fifties and sixties of the last century, 1 through the informative model (or customer model) prevailing in some countries in the seventies, 2 to the modern collaborative model of doctor-patient relationship that acknowledges a patient's right to fully participate in the medical decision process. These three basic models of doctor-patient relationshippaternalistic, informative and collaborativediffer in the perception of physician and patient roles, patient autonomy and the importance of the patient's values and preferences in the decision-making process. 3,4 Unlike the paternalistic relationship model, which assumes that the physician knows what is best for the patient, or the informative model, in which the decision-making process is transferred to the patient, 3 in the collaborative model, both doctor and patient are involved in Doctor-patient communication is a multidimensional process, which on the doctor's part includes delivering clear and complete information about the disease, its symptoms, possible treatments, their side-effects and the prognosis. 7,[36][37][38] In addition, doctor's assessment of the patient's understanding of the information is recommended, 27,39 as well as discussing patient's willingness to be involved in the decision-making process, 26,40 and support the patient emotionally. [41][42][43] Similarly, the patient is expected to provide complete information regarding his/her symptoms, 40,[44][45][46][47] ask questions regarding unclear information [45][46][47] and/or lacking information such as medical treatments and procedures, side-effects and prognosis. 40,45,48 In addition, expressing fears and worries by the patient is advisable. 18,41,49 In this model of interaction, the physician's knowledge of the patients' preferences and willingness to take part in the medical decision-making process are of special importance. 26,46,50 Several measures have been developed to evaluate physician communication patterns. 45,[51][52][53][54] As to the patients' communication behaviour, Galassi et al. 45 developed the Patient Reactions Assessment (PRA) which includes three subscales, two of them for evaluating the medical provider's contribution to the relationship, while the third subscalethe Patient Communication Index (PCI)was developed to assess only the patient's ability to initiate communication about his/her illness. Lerman et al. 46 developed the Patients' Perceived Involvement in Care Scale, which also includes three subscales: one measures physician behaviour, while the Patient Information Scale contains statements regarding patient level of information providing, symptoms and medical recommendation, and the Patient Decision-Making Scale which includes statements regarding the patient's suggestions and agreement to medical tests and treatment. However, certain aspects of care and patient communication, such as side-effects, prognosis, quality of life, fears and other-related emotions, are lacking in the previously developed scales. These aspects, which are important for evaluating the patient's behaviour in the physicianpatient interaction, have been included in the PCPS.

Patient Communication Pattern Scale (PCPS)
Considering the changes in the patient's rights and role in the doctor-patient interaction, and the absence of a complete measure to evaluate patient communication patterns, we developed the PCPS based on the above-mentioned communication literature, a qualitative study conducted on 12 oncologists in a number of medical centres in central Israel to understand how they perceive patients' preferences and behaviour in the doctor-patient interactions. The qualitative study supports the literature and suggests that only partial information is provided to patients, especially with respect to treatment options, side-effects and possibilities of recovery. The oncologists claim that patients primarily ask questions about side-effects and do not ask about prognosis; patient's preferences are hardly discussed and the treatments fit physician recommendations based on their experience and preferences rather than those of the patient. Conceptually, the PCPS was designed to encompass five different facets of communication: (1) relaying clear information about the illness and symptoms; (2) questioning and requesting clarifications; (3) initiating request for information from the physician; (4) guiding the physician according to one's own preferences; (5) reporting one's feelings.

Scale validation
Age, education, religiosity and desirability of control (in general and over the patient's own health in particular) were used to validate the scale. The literature indicates that elderly patients are more passive in doctor-patient interactions than younger patients, 49,55,56 tend to ask less questions, 47,48 do not present their feelings as much as others 47 and are more likely to leave the treatment decisions to their physician. 57,58 Studies also show that educated patients tend to ask more questions than less educated patients 48 and are more willing to participate in decision-making processes concerning their health. 58 Religiosity was also found to be related to the level of patients' preference for a collaborative model of doctor-patient interaction. For example, religious Israeli Jewish patients wanted less open communication with their physician about end-of-life care than secular patients, 59 and American religious individuals had higher levels of trust in their physicians than the less religious. 60 Consequently, we hypothesized that patients' age and religiosity would be negatively associated with patient's scores on the PCPS, while patients' education would positively correlate with their PCPS scores. Desirability of control reflects the degree to which a person is motivated to control his/her life's events. 61 Desirability of control (in general and over the patient's own health in particular) was assumed to positively correlate with patients' PCPS scores, as seeking as much information as possible is often the way of people to cope with uncertainty and gain more control over their lives. 62,63 Methods

Participants and procedure
The PCPS statements were formulated based on the literature, a qualitative study with oncologists, and a pre-test conducted on 25 patients waiting for appointments to their family physicians in community clinics. The interviews were approved by the Helsinki Committees of the Clalit Health Services (the largest HMO in Israel). The study for testing the PCPS was performed in an oncology day-care clinic of one of the largest medical centres in Israel that receives about 3000 new patients every year. Study procedures were approved by the Helsinki Committees of this medical centre which belongs to the Clalit Health Services. Patients who visited the oncology clinic for consultation and/or treatment and were Hebrew speaking and cognitively competent and able to independently answer questions, were asked to participate in the study. Those who agreed to participate in the study signed consent forms. Altogether, 251 adult patients with cancer participated in the study, 140 women and 111 men. The average age of the participants was 60.7 (SD = 3.1) ranging from 22 to 88 years; 247 (98.4%) of the patients were Jewish, 78.1% were married and 35.5% had an academic degree (Table 1). Age, education, religiosity and desirability of control Agepatients were asked to respond in what year they were born, and we calculated their age accordingly. Education was assessed by responses to being told to choose one of five levels of education: up to 8 years, partial high school, high school, tertiary and academic studies (see Table 1). Religiosity was measured by a short version of the Jewish Religiosity Scale; the score is the sum of five items: strength of belief, self-perceived religiosity, level of religiosity while patient was growing up, attending religious ceremonies and adherence to Jewish dietary laws, 64 with higher scores presenting higher levels of  61 . Desirability of control over the patient's own health was measured by one item: 'Regarding health matters, I've always preferred that the doctor decides what is best for me."; the patient was asked to express agreement on a 6-point scale, ranging from 1completely incorrect to 6completely correct (the answers were reversed, so that a higher score indicates a higher desirability of control over the patient's own health).

Statistical analysis
After conducting a factor analysis which revealed 5 facets, the structure of the PCPS was evaluated using confirmatory factor analysis (CFA), with AMOS 18.0. To validate the PCPS, we used Pearson correlation coefficients for evaluating the associations between PCPS and age, religiosity and desirability of control, and Spearman correlation coefficients for evaluating the correlations between PCPS and education.

The PCPS
Average scores and internal consistency of the indices are presented in Table 2. The factor analysis resulted in five factors. The general PCPS scores of our participants were relatively low with an average of 3.05 (SD = 0.85). When referring to the different facets of communication (Table 3) Confirmatory Factor Analysis (CFA) CFA was used to confirm an a priori hypothesis regarding items which should be grouped together. 65 The chi-squared, comparative fit indices (CFI), root mean square error of approximation (RMSEA) and HOELTER were used for evaluating and comparing models across the CFA. 66,67 After excluding all missing cases, responses of 223 patients were examined by CFA. In the first CFA model, the Information factor's loading was very low (0.16), but theoretically 'relaying clear information about the illness and symptoms' is part of the communication behaviour of the patient; additionally, taking out this factor worsened all the fit indices. The recommendation of the modifications in CFA included correlations between statements 8 and 13, and 5 and 10 (reversed); adding these correlations (one by one) to the CFA model improved all of the indices, but at the same time, adding the correlation   (Fig. 1).
Validity was assessed by correlations between the PCPS and age, education, religiosity and desirability of control in general and over the patient's own health. Although all of these associations were found to be statistically significant, some of them were relatively low. Therefore, we  The PCPS described in this study was developed based on a literature review, a qualitative study among oncologists who were asked to describe their patients' communication patterns, and a pre-test conducted on Israeli patients with cancer. The psychometric characteristics of the scale were assessed in a study conducted on 251 patients with cancer. Our findings support the a priori 5-facet conceptual structure of the PCPS. All of the indices we used indicated good fit of the model to the collected data.
Regarding the validity of the PCPS, as expected, univariate analyses resulted in statistically significant negative correlations between PCPS scores and both age and religiosity, and significant positive correlations between the PCPS scores and education and desirability of control scores (both general and over the patient's own health in particular). However, only age, education and desirability of control in general were found to be significantly explanatory factors of the variability in PCPS in the expected directions. In addition to supporting the validity of the scale, these findings indicate that the younger patients are and the higher is their education and need for control, the more they use open and collaborative communication patterns in doctor-patient interactions.
As regard to patient communication patterns, our findings indicate partial openness and collaboration among Israeli patients with cancer. It appears that patients relay information about their physical symptoms and ask questions about the information that is provided, but almost never request any further information. Patients rarely ask about other or additional possible treatments, side-effects or prognosis for recovery. Generally, patients also avoid initiating discussion about personal preferences and feelings.
When comparing the results of this study to previous reports, our findings also indicate a significant discrepancy between patient desires and behaviour. Generally, when asked, patients express a desire to receive as much information as possible about their diseases, 57,68-70 and to take part in treatment decisions. 40,58,71 However, there is hardly any evidence that these desires are acted upon in their behaviours within the patient-physician interaction, in support of previous findings of studies conducted on doctors and patients in Israel and other countries. 22,24,47,49 Our findings should be viewed in the light of the limitations of the current study: the internal consistency of two of our subscalesinitiating request for information from the physician (Initiation) and guiding the physician according to one's own preferences (Preferences)is low to moderate. Also, the general scale includes only one statement that is negatively phrased; this item has a low loading in the final CFA, which might be explained by its' negative wording. In addition, this study was conducted on patients with cancer in Israel. Therefore, further research to be conducted in other countries and on patients with other diseases is needed in order to broaden our understanding of patients' communication patterns.
Aside from its limitations, this study leads to the conclusion that in spite of current recommendations for open communication and active patient participation in the medical decision process, 41,72 it seems that patients have difficulties in exercising their wishes and rights, or may not be willing to know more about their diseases, especially when suffering from severe diseases such as cancer. To promote the collaborative model of doctor-patient interaction, it is desirable that physician enhance their communication skills, including techniques for eliciting patient preferences and encouraging participation in the decision-making process.
Regarding the PCPS, it was found to be a valid and useful tool for evaluating and understanding patient communication patterns, and for identifying the facets, which are not openly addressed. We are becoming more aware of the importance of open and collaborative doctor-patient communication, especially when patients suffer from severe and/or terminal diseases. From a practical point of view, the PCPS can be used for evaluating the level of patients' communication pattern on each of the scale's five dimensions. Such evaluations can lead further to the development of effective interventions for empowering patients to become more active participants in the medical decisionmaking process.

Funding
No source of funds.