Patient satisfaction with doctor-patient interaction and its association with modifiable cardiovascular risk factors among moderately-high risk patients in primary healthcare

Study design and sample

A six-month prospective survey was conducted among moderately-high CV risk patients in a primary outpatient care of a tertiary hospital in Kota Bharu from 2012 to 2013. We included: moderately-high risk patients with multiple (2 or more) risk factors and a 10-year risk for CHD of 10–20% (National Institutes of Health, 2001). The risk factors are cigarette smoking, hypertension (BP ≥ 140/90 mmHg or on antihypertensive medication), low-density lipoprotein cholesterol (LDL) (<1.1 mmol/L), family history of premature CHD (CHD in a male first-degree relative at <55 years, or in a female first-degree relative at <65 years) and age ≥45 years for men and ≥55 years for women. We excluded illiterate patients and those with established CHD or a history of stroke.

Sample size calculation was based on the first objective i.e., patient satisfaction on doctor-patient interaction. Based on single-proportion formula, the proportion of patients intending to adhere to treatment, a component of patient satisfaction, was 82.7% (Abioye Kuteyi et al., 2010). Using the precision of 0.08 and 95% confidence, the minimum required sample size was 86; after considering the non-response rate of 20%, the calculated sample size was determined to be 104. All of these patients fulfilling the eligibility criteria were invited to participate in the study.

Ethical consideration

The study protocol was approved by the Human Research Ethics Committee, Universiti Sains Malaysia (USMKK/PPP/JePeM[245.3.(9)]) on 6 February 2012. Patients were provided information related to the study along with an Information and Consent Form. Confidentiality was ensured and written informed consent was obtained.

Data collection

Patients attending the primary healthcare clinic were screened for the cardiovascular risk factors based on their medical records. Risk factors were calculated using the Framingham risk factors point score and 10-year specific risk by gender. These Framingham risk scores, delineated in Adult Treatment Panel III (ATP III), have been used since 2001 and consist of five risk scores (National Institutes of Health, 2001).

Prior to the day of the visit, patients were asked to fast overnight. On the day of the visit, 3 ml of blood were taken from each patient and sent for laboratory testing to determine fasting cholesterol levels. Also obtained during the visit were body mass index (BMI) (kg/m ²), waist circumference (WC) (cm), and blood pressure (mm Hg; measured with a calibrated standard digital blood pressure machine (Omron Global, Kyoto, Japan).

A structured intervention counselling with a standardize content for CVD prevention was carried out by the researcher; this ensured that there was limited bias in measuring patient satisfaction with the doctor-patient interaction. After counselling, patients were then required to answer the Malay version of the MISS-21 questionnaire outside the consultation room immediately after counselling. Upon completion, patients submitted the questionnaire to the research assistant. Six months after their initial visits, patients returned for follow-up appointments with the same healthcare practitioner, during which 3 ml of blood were drawn to obtain a fasting lipid profile (TC, HDL, LDL); blood pressure was also measured.

Research tool

Skala Kepuasan Interaksi Perubatan (SKIP-11) is a modification of the Malay version of MISS-21 suggested that the three-factor model with 11 items, acceptable for use in assessing patient satisfaction with patient-physician interaction in a primary healthcare setting because it is valid, reliable and simple. The three domains are as follows: (i) information provision by the doctor (distress relief; four items), (ii) patient’s confidence in doctor (rapport; four items) and (iii) doctor-patient interaction outcome (interaction outcome; three items). Cronbach’s alpha was 0.513 for distress relief, 0.708 for rapport and 0.747 for interaction outcome. The overall Cronbach’s alpha was 0.669, indicating a satisfactory level of internal consistency. The patients’ subscale scores were added together and the mean subscale scores were determined (Abd Aziz et al., 2013). Categorisation into satisfied and dissatisfied categories was based on the overall items and items for each subscale. Score >44 for overall items and score >16 for distress relief, >16 for rapport and >12 for interaction outcome subscales are considered as satisfied.

CVD risk prevention counselling applied a risk-stratification approach that covered adherence to medication and advice on smoking cessation, healthy dietary choices, physical activity and weight reduction based on recommended steps to help prevent cardiovascular disease by the WHO (2007).

Framingham cardiovascular risk assessment provides separate score sheets for men and women. The scores for age, total cholesterol, high density lipoprotein cholesterol, blood pressure and cigarette smoking are summed up and the 10-year coronary risk is determined. Patients are then divided into those with 10-year risk for CHD of >20%, 10–20%, and <10%. Moderately-high risk is used to describe individuals with multiple (two or more) risk factors and a 10-year risk for CHD of 10–20% (National Institutes of Health, 2001).

Data analyses

Data checking and cleaning were performed prior to data entry and analysis using SPSS version 19. Descriptive analysis was used to determine the overall patient satisfaction score and subscale of distress relief, confidence rapport and doctor-patient interaction outcome. Paired t test were used to determine the change in modifiable cardiovascular risk factors i.e., TC (mmol/L), HDL (mmol/L), and systolic blood pressure (SBP) (mmHg) that were based on risk factors used in the original Framingham scoring system (National Institutes of Health, 2001) between the appointments at baseline and six-month follow-up. Simple and Multiple linear regression confirmatory analyses were used to determine the association between patient satisfaction scores and modifiable risk factors (TC, HDL, SBP). Score changes in TC, HDL, and SBP served as the outcome variables. The fixed factor was patient satisfaction score; the adjusted variables were sociodemographic (age, gender, level of education) and medical (smoking status; presence of diabetes; presence of hypertension; baseline BMI, WC, SBP, TC, and HDL). General linear regression confirmatory analysis was used to examine the relationship between one numerical dependent variable and more than one independent variable with the goal to develop a best fitting, parsimonious, biologically sound and easy model to estimate the beta.

Patients’ satisfaction of the relation with their doctor

Using the ‘Skala Kepuasan Interaksi Perubatan-11’ Malay version of MISS-21, our findings revealed that about three-quarters of the patients were satisfied with their experience in primary healthcare clinic with their doctor-patient interaction. It was consistent with other studies that evaluated patient satisfaction in primary healthcare clinic, where 63.3% were satisfied in a study in Nigeria (Abioye Kuteyi et al., 2010), 74% in Trinidad and Tobago (Singh, Haqq & Mustapha, 1999) and 81% in the Netherlands (Van Uden et al., 2005). With regards to method of assessment for patients’ satisfaction, only Abioye Kuteyi et al. (2010) used MISS. Singh, Haqq & Mustapha (1999) used self-developed questionnaire to assess patient satisfaction with three groups of healthcare workers, namely doctors, nurses and pharmacists. About 74% of the patients were satisfied with doctors (Singh, Haqq & Mustapha, 1999), which was comparable with the present study. In another study, the CSQ was mailed to patients three weeks after consultation followed by telephone interview by the nurses (Van Uden et al., 2005). The satisfaction concerning visits to general practice (81%) was available in only one study (Van Uden et al., 2005) and the findings were comparable.

About 91% of our study patients were satisfied with the physician confidence rapport domain and 85% were satisfied with the distress relief domain; this was in line with a previous review (Williams, Weinman & Dale, 1998). Our major findings showed that the domain of distress relief was favourable among patients, and a good rapport was observed between the doctor and patient. Patient satisfaction was associated with doctor-patient interaction, specifically the patient’s confidence, perception of the doctor’s communication skills and perception of information provision. Meanwhile, adherence intent was related with patient’s confidence and perception of information provision by the doctor (Abioye Kuteyi et al., 2010).

Another interesting finding, about 24% of our study patients were not satisfied with the doctor-patient interaction using similar cut-off score for satisfaction. In other studies, this rate were 26% (Singh, Haqq & Mustapha, 1999) and 36% (Abioye Kuteyi et al., 2010) of the patients were not satisfied with the doctor-patient interaction. One study reported satisfaction as mean scores with higher scores indicating more satisfied (Van Uden et al., 2005).

Among our patients, 33% of them were not satisfied with the interaction outcome domain and, as a result, felt uneasy or unsure about following the advice given by the physician. The causes of unfavourable findings in this domain could be due to patients’ unwillingness to change their behaviour and adhere to the advice. A study by Berra (2003) on the Health Education and Risk Reduction Training Program (HEAR2T), which combined intensive behavioural risk factor control and medical therapies, showed that patient satisfaction towards the counselling given was very good (97%). However, only 69% of the patients that reported they were satisfied made behavioural changes, and 73% making behavioural changes expressed that they were confident in their ability to maintain these changes or adherence (Berra, 2003).

Association between patients’ satisfaction and cardiovascular risk factors

After six-month follow-up, our primary prevention counselling showed a relevant beneficial decrease in various modifiable cardiovascular risk factors (namely, TC and SBP). But reductions in SBP and TC were not related to the level of patient satisfaction and did not contribute to patient outcomes.

Satisfaction with the patient-physician interaction on the part of the patient does not signify that the patient will adhere to a medication; the converse is also true. Satisfied patients (63.3%) were more likely to adhere to advice from physicians (94.2%) but 56.6% of unsatisfied patients (36.7%) also intended to follow the advice (Abioye Kuteyi et al., 2010). In the absence of satisfaction, the tendency to follow advice (even in instances of dissatisfaction) persisted.

One study showed that physicians who demonstrated ‘attentive and respectful listening’ were able to give the perception that they ‘really cared’ for their patients; this perception also resulted in patient satisfaction (Jagosh et al., 2011). A Canadian longitudinal cohort study regarding patient satisfaction and its relationship with quality and outcomes of care after acute myocardial infarction showed that 91.7% of patients reported satisfaction with their overall care. However, this satisfaction was not associated with quality indicators for myocardial infarction and clinical outcomes (Lee et al., 2008).

Limitations and further researches

The strength and limitations of the survey reside in some points. The present study has few limitations. First, our study sample is limited to non-probability sampling and was performed in an outpatient clinic at a university hospital, and thus, does not represent all primary healthcare clinics. Second, the six-month duration from baseline to follow-up is short.

While the results are statistically significant in terms of reductions in cardiovascular risk factors (if not with respect to the association between these declines and satisfaction with physician-patient interactions), the long-term results concerning whether target levels for BP, HDL or TC were attained cannot be assessed. The study was performed in an outpatient clinic at a university hospital, and thus, does not represent all primary healthcare clinics.