Factors affecting patients’ ratings of health-care satisfaction

Patient surveys are commonly used in Western countries to evaluate patients’ experiences. However, to achieve the political goal of higher levels of patient satisfaction and greater patient influence on professional care, it is crucial to understand the motives involved in patients’ ratings [1-4].

Surveys often include a Likert response scale. Response categories may, e.g., range from very good to very bad or from agree to disagree [3, 5]. Thus, a Likert scale presents a way to capture the relative intensity of the patients’ emotions. When results of surveys with closed response categories were supplemented with subsequent semi-structured interviews, it became clear that questionnaire responses were too positive relative to the interview results [6]. These studies, however, were conducted between ten and 20 years ago and only in England. Additional space for comments is now generally used [6]. In the present study, which is based on a Danish questionnaire, the aim was to find out how patients associated their experiences of care which was rated on a Likert scale with the categories very bad, bad, good and very good.

METHODS

Patients’ references for ratings in a Danish national questionnaire were explored employing a phenomenological-hermeneutical approach. Participants were adult outpatients who had recently been discharged from a department for women’s conditions in a public Danish university hospital. Healthy women giving birth were
excluded.

The aim of the questionnaire [5] was to measure patients’ experiences of the health-care system and included space for comments. For our research purpose, questions about the patient’s general practitioner and overall experiences were omitted from the national survey [5]. This article focuses on the ten questions that required responses on a scale ranging from very good to very bad, or from very well to very poorly. In Danish language, the latter scale also translates into the equivalent of very good to very bad; therefore, we will refer only to the very good to very bad scale in this article. These questions addressed issues relating to information, reception, staff preparedness, collaboration between care units, treatment process and management, and the
handling of errors. Patients completed questionnaires during two randomly chosen weeks in 2011 and placed them in one of several mailboxes on the ward. In order to reduce recall bias and increase diversity in our subsequent sample, the mailboxes were emptied repeatedly, and patients who agreed to participate in an interview were divided into groups according to the care unit they visited. Patients were chosen randomly within these groups and many were interviewed [7] before they left the hospital. Additionally, in order to further verify our results in our own context, we continued to conduct as many interviews as possible for another six weeks.
We preferred face-to-face interviewing in a quiet setting, but telephone interviews were accepted on the patient’s request. Moreover, we also accepted if patients preferred to be interviewed without being recorded, and their validation of the interview notes was subsequently obtained within ten days. The patient’s completed questionnaire served as the starting point of the interview, followed by prompt questions (Figure 1). Fully recorded and transcribed interviews/patient-validated notes and the accompanying completed questionnaire were analysed and interpreted using an abductive process inspired by Paul Ricoeur [8, 9]. The process includes the following levels: 1) an overall, inductive interpretation in which all data were read as one whole text (a naïve reading). This led to 2) a structural analysis, which ran in a dialectical fashion between what was said (the actual words in the text) and what was talked about
(a structure of units of significance) and included only the interviews from the first two weeks. This process allowed themes to be generated. Thereafter, 3) a deductive process was applied in which the convergence between the themes and the whole text was investigated. These three steps were repeated until the structural analysis validated a naïve reading without contradictions, and before the themes were discussed in relation to relevant literature.

Ethics

Patients were informed verbally about the study, either by secretaries or by the first author. They were given a blank questionnaire, an information sheet and an invitation to participate in both the survey and in an interview about their experiences at the hospital. The study adheres to the Declaration of Helsinki [10]. The Danish Scientific Ethical Committees deemed it unnecessary to be involved in this project.

Trial registration: The study’s Danish Data Protection Agency number is 2008-58-0035.

RESULTS

Participants aged between 21 and 80 years completed the questionnaire. In total, 202 questionnaires were returned, and 19 patients were interviewed within the two survey weeks (duration 15-17 minutes). Additionally, we conducted 44 interviews within six weeks (average duration 16 minutes) (Figure 2). In the analysis and interpretation, two key themes were identified that relate to patients’ rating of their health-care experience: 1) the quality of the clinical service, which comprised both timeliness, and expected and experienced professional service, but not the professionals’ attitude and, 2) the health-care professionals’ attitude where key factors were the patients’ experience of recognition and respect. Recognition and respect encompass feelings of being acknowledged, listened to, heard and accommodated, taken seriously, valued, and that nothing is done that would be upsetting. The themes are outlined in
Table 1 and will be elaborated upon in the following
section.

The quality of the clinical service

Exemplary clinical service was experienced when professional services were received in a timely fashion and at the expected level. This was experienced reported by those who rated the care as very good. The rating “good” related to two different situations: either an exemplary clinical service or a dip in exemplary clinical service. An example of the former is the case in which a physician was not aware of being in earshot of the waiting patient at the outpatient clinic:

“The reception was good. The secretary was nice and quickly told me what I needed, and the physician smiled at me and shook hands with me. But I heard the physician I should have met say that he didn´t have the time for an examination and asked his colleague to do it. I felt like a number”.

An example of a dip in exemplary clinical service is the case where a patient met an unprepared physician, but the patient valued his preparation effort as good:

“Physicians should be prepared, but mine did not know me or my condition. That is bad, but a nice nurse took the time to tell me beforehand that several physicians were on sick leave. I felt it was OK in that case that the physician simply asked me about my condition. That was the quick way to get prepared”.

Patients’ expectations of clinical service included, e.g., being registered and informed swiftly so that they could prepare themselves for the upcoming situations, and receiving information when they asked for it, at the latest. They expected professionals to read up on their case before a meeting and to manage the course of treatment and care without errors, delay or loss of information. In situations that were judged as bad, the clinical service was not offered in a timely fashion, but still it was offered; and in very bad situations, patients left the hospital without faith that they would receive the service in question. For further examples, see Table 1.

As shown in both Table 1 and in the above-mentioned examples, patients combined the experience of clinical service and health-care professional attitude in their ratings.

The attitude of health-care professionals

A very good experience required that exemplary clinical service was provided with an attitude of recognition and respect. Experiences rated as good could reflect that there was a partial lack of this attitude, and recognition and respect were always partly lacking in bad experiences. Examples include situations in which the staff member was not the first person to greet the patient, did not fulfil promises or used technical terms that were hard to understand. Patients disliked when health-care professionals talked about them in the third person in their presence or verbally reduced them to a disease or an examination. In very bad experiences, patients felt that they were treated with indifference. For examples, see above and Table 1.

DISCUSSION

Our findings may be summarised as follows: when the patients rated the care they received, they evaluated experienced care against expected care. Therefore, their rating also reflected their relations with the health-care professionals. Our results identify two types of relation of significance for patients: The patient-professional relation (a care relation) in which clinical service was important, and a human relation in which attitude was important. It appeared that, for the patients in our study, an acknowledging and respecting professional (acting as co-creator of a human relation) had the power to improve patients’ experiences, a factor that was also reported by others, e.g. [11, 12], and found in relation to the Likert scale rating from good to very good [13]. This contradicts other research findings where the possibility of raising a rating from good to very good is questioned, e.g. [14]. However, our results also supplement that of others who succeeded in mapping only 17.5% explanatory factors for patient satisfaction and did not include health-care professionals’ attitudes [15]. Our results may suggest that excellent clinical quality could also improve patients’ rating at the middle of the Lickert scale. At the low end, it was the care relation that proved to be the more important, because the human relation could not compensate if patients were not allowed to have a patient role (i.e. left the hospital without faith in getting the care they needed). Clinical service and acknowledging and attentive health-care professionals have been found to be important for patients for many years, e.g. [16, 17], but the way in which patients may combine these two factors in relation to a Likert scale has not been investigated before. To put our findings into a wider context, we sought a theory of human relations that could explain our findings. Løgstrup states that, basically, we all seek to have good, trusting relations. This is achieved if the other party behaves as we expect. However, if the other person does not conform to our expectations, we need to take into account a greater or lesser degree of disappointment.

This could lead us to regard the other party in a negative way [18]. In other words, if health-care professionals deeply disappoint a patient, this patient may interpret signals from these professionals, or all, health-care professionals in a negative way; e.g., subsequent attention and smiling may be interpreted as “fake”. In relation to our results, patients rewarded the health-care professionals who behaved as expected. Moreover, none of our patients gave information about helpful human relations in situations where they experienced a total lack of clinical quality; on the contrary, they felt that they were being treated with indifference. Løgstrup’s theory deals with trust and distrust in human relations. He specifies that trust is an ethical requirement in all human interaction and considers the theory to be universal [18]. The fact that this theory fits well with our data may indicate that trust (or its lack) could be one of the factors underlying a choice of rating. In that sense, our findings support the idea that trust and patients’ grading of experienced care are interrelated [19]. Such relations will need further investigation.

Limitations

The strength of this study is its combination of qualitative interviews with a Likert scale survey which allowed for an investigation of how patients associated their experiences to the rating scale. The trustworthiness of our data could be questioned because the time from questionnaire completion to interview was up to six weeks and the length of the interviews was short in terms of qualitative methodology. Moreover, we used several
data-generating methods of which telephone interviews lacked the visual element. However, for the initial 19 patients, all interviews were face-to-face, the time between events was a maximum of two days and our data collection was focused and included many parallels drawn between the individual answers. We gathered rich descriptions and considerations; albeit the final
interview was not as detailed as the others. We included data from many patients which may induce cursory analysis and interpretation [7]. However, the second level of the analysis, which gave rise to themes, was based on the first 19 interviews, and this ensured overview and analytical depth. We experienced data saturation within the initial 19 interviews, but were able to provide further verification in the data from the rest of our participants. Furthermore, the deductive analysis did not reveal contradictions, and this strengthens the internal validity of our analyses. Our participants were adults (21-80 years of age) and they visited different wards and outpatient clinics; however, only women were included.

CONCLUSIONS

Our female patients could associate their experiences with their ratings, and two types of relations seemed to be at play: a care relation and a human relation. Within these relations, the clinical service provided and the attitude of health-care professionals were important. Patients combined their experiences from these two factors when they evaluated the care they received.

Care rated as very good represented excellent clinical service provided with recognition and respect, whereas a good care rating reflected a dip in either one of these factors. Our results may inform health-care practice, but department-specific examples may be needed to initiate improvements.

Correspondence: Marianne K. Thygesen. E-mail: marianne.thygesen@health.sdu.dk

Accepted: 21 August 2015

Conflicts of interest:Disclosure forms provided by the authors are available with the full text of this article at www.danmedj.dk