Same questions, both sides

Interpersonal drivers of compliance · MSc thesis, University of the Aegean · 2018–2020 Fieldwork and analysis, under my supervisor’s guidance 115 doctors · 65 patients · 60 medical associations contacted · 7 hypotheses · SPSS

---

The question

Patients often don’t follow medical advice: the medicine, the diet, the change in routine. My first design MSc ended with a thesis on why, approached from one angle: the relationship between doctor and patient. How a doctor asks, how a patient sees the doctor, and which factors go with a patient actually doing what was asked.

The topic, the research model and the measurement scales came from my supervisor. I ran the study and did the analysis, and when I got stuck I went to her for guidance.

Scales borrowed from somewhere else

The scales didn’t come from medicine. They came from business-to-business marketing, where researchers measure how one company gets another to comply: a supplier and a retailer, a manufacturer and a distributor.

• Influence strategies (Payan & McFarland): six ways of asking. Giving reasons, plain requests, sharing information, threats, promises, recommendations.
• Compliance (Hunt and colleagues): 14 questions on how likely the other side is to go along.
• Role conflict and role ambiguity (Singh; Rizzo and colleagues): whether the doctor knows what the job expects of them.
• Self-identity, loyalty and satisfaction (Jamal & Adelowore).

Then seven single questions for factors the medical literature links to compliance: prior trust, prior satisfaction, loyalty, how serious the condition is, the patient’s age, how much public attention the condition gets, and the cost of the treatment.

Doctors and patients got the same scales, each from their own side of the consultation: doctors answered about “the majority of my patients”, patients about “my doctor”. The thesis notes that, as far as was known, this combination of scales had not been measured in Greece before.

Reaching doctors and patients

I contacted all 60 medical associations in Greece, some by phone, some by email, with the questionnaire for them to forward to their members. Some agreed straight away, some put it to their board, and the two largest, Athens and Thessaloniki, said no.

The patient questionnaire ran for ten days on a Rhodes local news site, in Facebook groups, and on the health site iatronet. Doctor Anytime never replied.

Both questionnaires were piloted first with 3 doctors and 8 patients, checking how long it took, which questions people struggled with, and whether the online form worked. The pilot cut the scales down, reworded the items people found hard to answer, and changed the research model and the hypotheses.

115 doctors answered, November–December 2019, from more than twenty specialties. A third had practised for over twenty years. 65 patients answered, mostly from Rhodes and Athens. Every question was mandatory, so there were no gaps in the data.

What the study found

Each hypothesis was tested as a Pearson correlation between one compliance question and one question about the factor, separately for each sample.

n.s. = not significant. Every other value is significant at the .01 level.

Both samples linked compliance to the relationship: satisfaction, trust and loyalty, along with how serious the condition is. For the doctors, satisfaction was the strongest link.

They parted on three factors. For the doctors, the patient’s age, how much public attention the condition gets, and the cost of the treatment showed no link with compliance. For the patients, all three did.

I also ran reliability checks on every scale (Cronbach’s α above .7 on all of them, in both samples) and an exploratory factor analysis on each sample.

What it can’t tell you

• Neither sample is big enough to generalise from. The two largest medical associations are missing, and the patients come mainly from two places, Athens and Rhodes.
• Correlation is not cause. The thesis ends by proposing a causal study of satisfaction next, because it stood out.
• One question per factor. Each hypothesis rests on a single pair of questions, not on a full scale.

After

The thesis closes by naming the next field worth studying: compliance supported by digital tools, apps and devices built to help patients follow advice.

My second MSc, two years later, ended in a pharmacy app: Pilly.