Utility of qualitative methods in a clinical setting: perinatal care in the Western province

Objective A peculiar paradox that has been observed in previous studies of antenatal care is where patients are satisfied with the services despite obvious lack of basic facilities. Qualitative methods were used to describe the experience of perinatal care in the Western province with the objective of demonstrating application of this method in a clinical setting. Methods This paper used a ‘naturalistic’ approach of qualitative methods. In-depth interviews conducted with 20 postnatal mothers delivering in tertiary care institutions in the Western province was tape recorded, transcribed and content analysed. To ensure objectivity and validity of results, the principle investigator received only the anonymised data to prevent any prejudices or pre-conceptions affecting the results. Results The main themes emerging from the text demonstrated ‘naïve trust’ in the carer and a state of ‘hero worship’ where patients were distanced and therefore unable and unwilling to query the decisions made by the carers. This is similar to a state of patientcarer relationship described in a published model known as guarded alliance, where the relationship develops though four phases based on the level of trust and confidence in the relationship. This state explains not only why patients fail to recognise and report any deficiencies in the services but also the need for them to justify the behaviour of caregivers even when it amounts to incompetence and negligence. Conclusions Qualitative methods allow the researcher to capture experiences in its ‘natural’ form rather than based on pre-determined protocols or plans, which may be limited to our own understanding and expectations and therefore unable to explain many idiosyncrasies of the programmes. This paper argues favourably for the use of qualitative methods in other clinical settings. Ceylon Medical Journal 2012; 57: 10-13


Introduction
Qualitative methods remain uncharted territory among local medical researchers, as evident by its scarcity in published papers. Only a few purely qualitative papers can be retrieved in past issues of the Ceylon Medical Journal from the last 2 decades and even these limit the qualitative aspect to the data collection, steering well clear of qualitative analyses [1,2]. This paper is an attempt to demonstrate the usefulness of 'pure' qualitative methods in clinical care. The choice of setting was perinatal care; a service that continues to present challenges. Despite high quantitative coverage of skilled attendance at delivery, perinatal morbidity and mortality rates have remained stagnant in recent years [3].
Attempts have been made to measure quality of antenatal care in Sri Lanka using questionnaires and checklists [4,5]. They have shown that a majority (75 -90%) are satisfied with the services offered at antenatal clinics [4]. However, when service provision is assessed independently, there are many deficiencies: incompetent staff, lack of facilities, and inadequate supplies. Interestingly, the majority of mothers did not report these deficiencies, nor did it seem to affect their 'satisfaction' with the services with many confirming that they will return to the clinic for subsequent care. This creates an interesting paradox, where clients are satisfied with the services despite obvious deficiencies. It could be either that assessments based on client perceptions does not correlate with the actual level of quality or that the methodologies used are unable to fully capture the characteristics of the programme.
This paper aims to explore the yet underutilised qualitative methods, to describe perinatal care in Western province tertiary care hospitals to demonstrate the versatility of this method not only as a research tool but also as an aid in programme evaluation.

Methods
A descriptive cross-sectional study was carried out using qualitative methods (in-depth interviews) based on principles of naturalist inquiry where experiences are captured in its natural form [6]. The Western province was the study area of choice as it represents the most developed tertiary care units in the country that ensures basic facilities to all women included in the study. Vol. 57, No. 1, March 2012

Papers
Women attending tertiary care hospitals for normal vaginal delivery during January-June 2009 were chosen by purposive sampling. Those with complicated deliveries and/or antenatal periods were excluded. Purposeful sampling method was used to select information-rich cases for in-depth study. Twenty mothers were purposefully sampled from two settings; 12 were selected from four randomly selected tertiary care hospitals and another 8 who had been discharged after delivering in these institutions were interviewed at home, within 5 days of delivery.
Informed verbal consent was obtained from participants. They were only required to describe the experience during the hospital stay, not to assess the care given to them. Complete anonymity and confidentiality of information was assured. Permission was taken from all institutional heads to conduct the interviews and it was assured that persons or places would not be identified in reporting results. The Ethics Review Committee of the Faculty of Medical Sciences, University of Sri Jayewardenepura approved the study.
A research assistant (RA) trained by the principal investigator (PI) conducted the interviews using a flexible interview guide, lasting 30-45 minutes. There were no direct questions about the quality of the services received or the appropriateness of the management. All interviews were tape-recorded with the participants' permission and any reference to persons or places were removed from the recording. The RA transcribed the recordings to text, verbatim, and also included any non-verbal communications. Data were analysed by the PI and it was maintained in the original language; Sinhala, during the process of coding and condensation and was translated into English only in presentation of results.
The transcribed text was analysed using qualitative content analysis method where each interview was read through to identify parts of text that have meaning according to the objectives of the study [6,7]. Each section of text identified was a 'meaning unit' and these were given a code ( Table 1). The coding was independently verified by a colleague for consistency. The meaning unit can be used in a number of ways in qualitative analysis [7,8]. Alternatively, the codes are re-arranged to create new themes and these are discussed in relation to a known concept, theory or simily. An inductive process was used in this analysis, where the text was first read in whole, attempting to make sense of the situation [8]. Then the text was more closely examined for similarities and deviations which could be explained using known simily, concepts or theory.

Results
All twenty women sampled were postnatal mothers who had delivered without any complications. Their ages ranged from 19 -36 years. Six of them had given birth for the first time and the others had been through a previous delivery. All ethnic groups and social classes were represented.
When examining the meaning units of text, they seem to fit the characteristics of an existing model of doctor patient interactions -the guarded alliance model [8]. This is a model developed using qualitative methods, phenomenology and grounded theory. This describes four phases of a relationship between a patient and a care giver: hero worship, resignation, consumerism, and team playing ( Figure 1). According to the model, development of the patientcarer relationship is a balance between the degree of trust and the degree of competence in their roles for both the patient and the carer. The state of 'hero worship' is characterised by a high degree of trust in the carer, and patients lack competence in decision making. The state of resignation lacks trust as well as competence in decision making; consumerism is a state of distrust where patients feel the need to demand care, competent in their role as consumers. Team playing, the most favourable state is where both the patient and the carer are equal partners making collective decisions with highest degree of trust and competence.
In this study, two important themes emerged from the text, which could be described as 'the doctor knows best' and 'they are gods or like gods'. The underlying core variable in both these themes was naïve trust, an important element in the first state of the guarded alliance model [9]. These two themes are discussed below.

'The doctor knows best'
During the process of labour and delivery, it was evident, that mothers, even those who had undergone previous deliveries were not aware of what was happening or was going to happen to them. The mothers' accounts frequently contained statements such as 'I was scared/ worried/did not know what was happening', indicating their fears. There were indications of complete lack of awareness 'something was pulled/cut/stitched, etc.' However, none of these fears or worries had been expressed or indicated to the doctor. This is one of the main features of hero worship; the high degree of naive trust, believing wholly in the goodness of the doctors' decisions.
Caregivers as well as patients contribute to the state of their relationship [9]. In this study, the caregivers were in part responsible for maintaining distance and hence a state of unapproachable 'divinity'. Most decisions, especially those made by the senior doctors, were not directly communicated to the mothers; it was through a nurse, medical student or a junior doctor (in that order). This created an 'aura' surrounding the 'heroes' which discouraged mothers from questioning authority.
For example, a woman believed that she was wearing the vaginal speculum for 2 days (to help in the delivery) as she had not seen it being removed the first time, nor had it been explained to her, leading her to erroneously conclude that she was wearing it for two days.

'They are gods/ like gods'
At the end of the interview, the RA asked the mother whether she was satisfied with the services received. Even in this study, as in all previous quantitative surveys, all 20 mothers reported they were 'satisfied' with services and *Includes the roles of the patients as well as the caregiver

Hero Worship
Team player

Degree of trust in health professional
Degree of competence within the role* Figure 1. Relationship styles within the guarded alliance [9] Papers would also recommend it to others [4,5]. In a number of instances, the women actually used the word 'god' to describe the caregiver or equated them to 'god' during the narration. They were 'very pleased' with the outcome -their new born baby -and was therefore quick to forget and forgive past omissions ("sins") committed by them.
Therefore any survey merely posing the questionwere you satisfied with the services? -would only obtain positive answers when the relationship is defined by naive trust and caregivers are 'worshipped like gods'.

Discussion
In-depth interviews with the postnatal mothers based on naturalistic principles (7) gathered a rich and varied mass of information that adds a different perspective to a poorly understood phenomenon [6]. The versatility of the method is its ability to show unanticipated variations and important idiosyncrasies of the programme unlike the predetermined responses in a checklist or a questionnaire, which are limited by our own understanding, which may not be accurate.
One of the common criticisms of qualitative research is its limited capacity for generalisation as they are based on small samples. Qualitative data although lacking in statistical inference, provide depth and breadth so as to permit 'extrapolation' in similar settings [9]. This study was limited to the Western province and women underzgoing normal vaginal deliveries, and therefore applicable to situations under similar, although not identical conditions.
The objectivity and validity of qualitative methods have been challenged, and the most important element in this sense is the researcher -who should remain an impartial inquirer, not predisposed towards certain findings ahead of time [9]. The PI in the present study took every effort to ensure neutrality. The data collector was a sociology graduate, an 'outsider' with no particular interest in the results. The PI refrained from visiting any of the institutions surveyed and the data collected were anonymised so that it was not possible for the PI to identify any of the institutions or personnel in analysis.

Conclusion
Qualitative methods are being advocated as more suitable for evaluation of programmes that aim for an individualised outcome such as patient care services [10].
This study demonstrates its usefulness in addressing an important paradox in previous assessments of quality, explaining the underlying state of hero worship in doctorpatient interactions. Although the present analysis is limited to perinatal care it is unlikely to be exclusive to this setting and argues favourably for use of qualitative methods in evaluating other fields of clinical medicine.