Vijayaprasad Gopichandran: How can we measure patients’s trust in doctors?

vijayaprasadJum Nunnally, the much acclaimed author of “Psychometric Theory” the standard textbook of psychometrics, which has run into several volumes, says “an accurate method was available for measuring the circumference of the earth 2000 years before the first systematic measures of human ability were developed.”

He expresses surprise that psychometrics took so long to develop as a specialty. Scholars still debate on the scientific rigor of psychometrics and its validity in assessing ability. As a physician and public health researcher my forage into psychometrics is recent and I had a very fleeting affair with the field in my efforts to develop a scale to measure trust in doctors.

Initially I conducted in depth interviews among patients and people in rural areas to understand their concepts of trust in doctors. The qualitative enquiry helped me understand some nuances. I understood that in low and middle income settings with inequities in access to health and resource limitation, there are certain unique characteristics of trust in doctor-patient relationships. Communities in these low resource settings had certain perceptions of competence of the doctor which helped them articulate trust. Often these competence judgments were based on personal experiences or experiences of other community members whom they trust. Competence was judged based on experiences of cure as well as fulfillment of certain expectations from treatment such as injections, intravenous fluid infusion, etc. In the setting of lack of universal access and health security in these low resource areas, assurance of treatment was another important dimension of trust. Dependable treatment at any time of the day for emergencies, irrespective of ability to pay is a marker of trustworthiness. Once a community placed trust in a doctor they were willing to make several compromises. They were willing to accept rude behavior of the doctor, they were willing to wait for a prolonged period of time to see the doctor and in some instances were even willing to pay exorbitant amounts of money out of pocket to consult the doctor. They also placed a high level of loyalty on the doctor they trusted. They would repeatedly visit the same doctor for all health problems, take their friends and family to the doctor and consult the doctor before taking any health related or personal decisions. Trust in the doctor was also articulated as a high level of respect for the doctor. From the qualitative exploration I understood that the community trusts the doctor whom they perceive as competent and dependable and when they trust them they are willing to accept drawbacks, are loyal to them and respect them. More details of this study can be found in this paper.

From this understanding, I wanted to develop a scale to measure the level of trust that a patient has on their doctor. All I needed to do was to capture the same dimensions of competence, dependability, loyalty and respect through a questionnaire and validate it as a measure of trust. I had to covert this conceptual understanding of trust into a quantitative measure. I started by converting the dimensions into representative statements. The patient had to read these statements and tell me whether they agree or disagree with the statements and to what extent they agree or disagree. For example I would state “I think my doctor gives me the appropriate medicines for my illness” representing the competence dimension. After generating a total pool of 31 questions, I interviewed more than 600 people, randomly selected from the community. I analyzed the data I gathered using both classical test and item response theories which are two paradigms of psychometric evaluation. After this I found twelve questions which capture the trust in doctors in this community with reasonable validity and reliability in the medium to low trust range. The final list of statements which featured in the scale are shown in the adjoining table. Psychometric properties of the scale showed that it has good validity and was a reliable measure of trust. The findings of this psychometric scale development process are published in this paper.

The final list of items in the trust in doctors scale

  • The doctor does appropriate tests to diagnose my disease
  • The doctor gives appropriate medications for my disease
  • The doctor’s treatment relieves the illness quickly
  • There are no side effects to the medicines prescribed by the doctor
  • Friends, relatives and neighbours recommend the doctor
  • I am confident that my illness will improve when I go to the doctor
  • If I go to the doctor, I will receive good treatment for my illness
  • The main intention of the doctor is to treat my illness and not anything else
  • Whatever illness I have, I will go only to this doctor first
  • I will recommend this doctor to all those who ask me
  • I respect the doctor a lot
  • I think the doctor is a very learned person

I thought I had successfully converted the concepts into numbers. After repeated use of this scale, which has good psychometric properties, I have now started to struggle with moving back to the concept of trust from this scale. In one of my validation exercises I found that the scale accurately predicted a person with high trust in about 80% of the occasions. However, I am starting to question whether these 12 items will cover the entire construct of trust in doctors. Each day I talk to a different person with a new perspective of trust in doctors. Each clinic day throws open experiences which make me question my own trustworthiness. The way each patient interacts with me in my clinic is a new story of trust. I still do not know if I have managed to go back to the concept of trust from the psychometric exercise of developing a scale.

I have learnt a few important lessons from this experience. While quantification and measurement are very important for science, it has to be an iterative process where one is willing to go back and forth from concepts and numbers. All psychometric scales, however old, time tested and rigorously validated they may be, should be interpreted keeping in mind the limitation of attrition of meaning in conversion from concepts to numbers. Finally scientists from a positivist paradigm need to get out of our comfort zones in dealing with quantifiable measures and become more comfortable with abstract concepts and accept that they could be scientific too.

Vijayaprasad Gopichandran is assistant professor of community medicine at the ESIC Medical College and Post Graduate Institute of Medical Sciences and Research, Chennai, India. His research interest is in doctor-patient trust relationships.

Competing interests: None declared.