Client Satisfaction Exit Interviews: Assessing Quality of Public Health Institutions through Generated Feedback

Inter-linkages between clients' satisfaction, quality of services, access to services and sustainability of public institutions play vital roles in ensuring effective governance in today's world. For instance, dissatisfied clients will opt out of the services, spread messages about their negative experiences discouraging others to access the services. This in turn will result in the breakdown of public institutions. Clients' satisfaction is therefore the key to their decisions to continue accessing public services and is essential to long-term sustainability. Dissatisfaction among clients could be either due to poor quality of services like cumbersome processes, untrained/unhelpful personnel or lack of facilities because of issues related to accessibility—distance between the clients' place of residence to the Public Health³ Institution. Cumulatively, all these factors with a lack of opportunity to express their opinions will result in the disengagement of citizens from public institutions. Performance from clients' perspectives, it has been argued, is better measured through their satisfaction levels rather than institutional data.

As a part of a large-scale effort to reinvent healthcare institutions, client satisfaction surveys are increasingly being promoted as a method of understanding quality of health services and the needs and expectations of clients from these institutions. Evidence from healthcare research across the globe reveal that research techniques such as client exit surveys provide a quick and inexpensive way of determining specific areas of health services where quality could be improved. Exit interviews have become a popular way of assessing client satisfaction in developing countries like Ethiopia (Fekadu, Andualem, & Yohannes, 2011; Tesfaye, 2005), Madagascar (Glick, 2007) and Bangladesh (Aldana, Piechulek, & Al-Sabir, 2001). The Department of Community Medicine, Midnapore Medical College, West Bengal conducts institutionalized exit interviews so as to understand clients' satisfaction levels as an indicator of their performance (Das et al., 2010). These exit interviews are useful for highlighting the inadequacies in the basic aspects of service delivery that affect the quality of clients' experiences. They also provide evidence of whether specific policies have their desired effects. They are found to be easier to administer, less expensive, and provide rapid feedback when compared with client report cards. While exit interviews are being widely used in healthcare research, one must also be aware of the possible issues related to the reliability of the data collected through exit interviews. One of the biggest criticisms of exit interviews is the problem of ‘courtesy biases’ (Glick, 2007). Courtesy bias results from the reluctance of the respondents to express negative opinions of service while they are within the premises of the service provider during the exit interviews. This could result in over-reporting of satisfaction levels.

As satisfaction and its measurement are important for health planning, Crow et al. (2002) reviewed 139 articles on determinants of satisfaction. Their review provided evidences on two groups of factors: first, those relating to the characteristics of respondents and second, those relating to healthcare delivery. No firm conclusions can be drawn on the relationship between levels of satisfaction and gender (Crow et al., 2002). It has been argued that satisfaction is high when favourable experiences match favourable expectations and low when negative occurrence reinforces negative expectations (Linder & Stewart, 1986; Swan et al., 1985). Studies also confirm a positive relationship between satisfaction and previously recorded levels of satisfaction (Gray, 1980; John, 1992). In fact, evidence indicates that levels of satisfaction get enhanced when the patients' wishes are met and dissatisfaction arises when they are not (Jackson, Chamberlain, & Kroenke, 2001; Weingarten et al., 1995). Doctors' communication skills and the extent to which the patients are open to reasoned arguments also help in shaping the levels of satisfaction (De La Cuesta, 1997). Fitzpatrick (1991) and Pickering (1993) have advocated the need to educate consumers about appropriate care and to manage their expectations about what can realistically be provided.

There is evidence that poorer physical health (Annandale & Hunt, 1998; Hsieh & Kayle, 1991; Rogut, Newman & Cleary, 1996) and low quality of life (Weingarten et al., 1995) are associated with lower levels of satisfaction. Nonetheless, it has also been found that older respondents are significantly more satisfied than others (Al-Bashir & Armstrong, 1991; Cleary et al., 1992). In general, the findings about relationship between socio-economic indicators—education, social class, occupation and income—have been inconclusive and at times inconsistent ( Bertakis, Roter, & Putnam, 1991; Cohen, 1996; Crow et al., 2002; Khayat & Salter, 1994). It has also been suggested that a personalized approach of physician is appreciated by healthcare seekers (Hjortdahl & Laerum, 1992; Holloway, Matson, & Zismer, 1989). Studies investigating levels of satisfaction with in-hospital care have identified quality of patient –practitioner relationship (Cleary et al., 1992; Howie et al., 1998) and discussions with the practitioner (Calnan et al., 1994; Coyle, Calnan, & Williams, 1992; Roghmann, Hengst, & Zastowny, 1979; Snell, 1996; Williams & Calnan, 1991) as most important factors governing client satisfaction. While in-hospital satisfaction has been reported to be lower in government hospitals compared to private hospitals (Fleming, 1981), nursing strain and exhaustion have been found to adversely affect patient satisfaction (Leiter, Harvey, & Frizzell, 1998). On the other hand, studies on out-patient care suggest that professional standards and interpersonal relationships are of prime importance to patients (Bishop et al., 1991; De La Cuesta, 1997; Fitzpatrick, 1991). Longer consultation and personalized approaches are also appreciated by out-patients (Hjortdahl & Laerum, 1992). It has been argued that affective behaviour like warmth and respect (Kenny, 1995), courtesy and apathy (Comstock et al., 1982) and sensitivity and understanding (Treadway, 1983) are associated with higher consumer satisfaction.

Monitoring users' satisfaction is important for maintaining service quality in treatment (Crow et al., 2002). Moreover, these findings can affect health outcomes (Bishop et al., 1991; Fitzpatrick, 1991) as unsatisfied and frustrated patients may not respond fully to therapeutic interventions (Strasen, 1988). With respect to information exchange, evidence shows that satisfaction is higher with feedback and discussions about treatment (Brody et al., 1989; Hall, Rote, & Katz, 1988; Jackson et al., 2001). Despite these assertions, several problems arise while measuring satisfaction (Williams, 1994), which weaken the policy implications of evidence-based advocacy, emanating from these findings. Specifically, individual's judgements reflect their own circumstances and background, and do not necessarily lend themselves to objective measurement. If expressed satisfaction results from users' knowledge and expectations, the comments on quality of services may remain uncharted in these responses. Therefore, it is desirable to separate subjectivity, emanating from expectations, from the objectivity of service delivery quality. Crow et al. (2002) suggest that further research is warranted with respect to methods adopted especially on timing of the survey, bias introduced by interviewers and how these exit interviews may be incorporated into healthcare decision-making.

This article analyses the findings from the exit interviews with an objective to understand (i) how the perceptions are shaped; (ii) could the quality of healthcare infrastructure and services be taken seriously for advocacy; and (iii) feedback from health administration on the findings. This exercise will reveal whether the client satisfaction exit interviews are indicative of the institutional and procedural aspects of the public health institutions. Thus, exit interviews could be used as a practical way of investigating clients' satisfaction with respect to accessibility and quality of services.

The present study was spread across the different geographical regions of the state. It covered 12 districts where Samarthan (2011) works intensively on governance issues related to NREGA,⁴ NRHM,⁵ RTI,⁶ water and sanitation, and social security programmes through policy advocacy with district administration. The districts were grouped into five clusters for ease of data collection. These clusters included the economically backward regions of Bundelkhand (Panna, Satna and Rewa districts). Bundelkhand is the poorest region of the state with a very low level of institutional delivery. The second cluster was from south-western Madhya Pradesh (Dhar and Jhabua districts) having highest percentage of Scheduled Tribe population. The region has high rates of poverty in both urban as well as rural areas. Three districts were also taken from Chambal region of Madhya Pradesh (Gwalior, Bhind and Datia districts). This cluster with a concentration of Scheduled Caste population has the lowest sex ratio in the state. The Jabalpur cluster included Jabalpur, Damoh and Seoni districts. Sehore district, which was 45 km from the state capital, was also considered from the central part of the state. The coverage of exit interviews is given in Table 3.

In each of the 12 districts, 1 CHC was selected randomly. In this CHC, 5 PHCs were randomly selected. A total of 4,219 individuals (3,308 out-patients and 911 in-patients) were interviewed from the PHC, CHC, district hospitals and private clinics to understand their perception and experiences of healthcare services. The data was collected between July 2010 and August 2010. Respondents attending an out-patient consultation were randomly selected on attendance and requested to complete the exit interview questionnaire. The research was conducted through a structured exit interview schedule which was used to collect data on the socio-economic background of the respondents, experiences of using the public health institutions, and the levels of satisfaction with respect to the quality of services. The key method used to analyse the data is quantitative. Tabular analysis is used to understand the degree of association between socio-economic characteristics of respondents and the level of satisfaction reported by respondents. This identified the main factors that were used in analysis of variance (ANOVA), which explained variations in satisfaction levels, as a cause-and-effect method. Although the present analysis has heavily relied on ANOVA, there are some sophisticated techniques—like Logit analysis—to explain the probability of being satisfied rather than not owing to changes in characteristics of respondents, and logistic regression—that can be applied. As the data on the level of satisfaction is ordinal, it was thought appropriate to use Analysis of Variance to identify factors that explain variations in patient satisfaction. Moreover, the latter techniques have stringent assumptions, making the results difficult to interpret, if assumptions are not met (Field, 2009). On the other hand, the assumptions of ANOVA are robust and variations from assumption, in practice, do not significantly affect the results (Field, 2009). Nonetheless, it is worth mentioning that though the variables, hypotheses and relationships tested in the article have already been studied before (see, for details, the literature review), the analysis has elements that lead to insights for policy formulation.

In all, 4,219 patients were interviewed from 12 districts of Madhya Pradesh. The satisfaction levels of patients were recorded on three different issues: behaviour of health professionals (doctor, ANM, nursing staff, ward boy, compounder, laboratory staff and medicine distribution person); infrastructure (building, waiting room, general cleanliness, toilets, safe drinking water, bed, testing instrument, meals, ward cleanliness and services of health institutions); and services (including availability, quality and capability of doctor/ANM, ambulance, OPD timing, laboratory tests, free medicine and availability of medicines). In the present analysis, the aggregated and averaged score for individual respondents' satisfaction for the three independent variables (i) behaviour of health personnel; (ii) infrastructure and (iii) services provided by health institutions were used. Table 4 provides the level of satisfaction recorded by the patients relating to these three variables. The findings reveal that while 61.8 per cent of the patients are satisfied with the behaviour of health professionals, rest are either indifferent or are dissatisfied. In case of infrastructure, only 42.5 per cent of the patients are satisfied and the rest are either indifferent to the infrastructure or are dissatisfied. In the case of services provided by the health institutions, patient satisfaction is relatively low at 41.8 per cent.

Two important and inter-related questions that we encounter in clients' feedback are: Are these opinions and reported experiences a valid conclusion on quality of services? This is an issue of methods used in generating the data and consistency in the patterns of the reporting. The second issue is: If these results are acceptable, how to use these variations as an advocacy tool? Before turning to the second issue, it may be worth reminding ourselves that as the OPD patient-exit survey was conducted outside the institutions, the likelihood of ‘courtesy bias'was relatively low. Moreover, if the proportion of highly satisfied or highly dissatisfied patients is less than 3 per cent and those who are not satisfied are about 42 to 57 per cent, the pattern of responses seems to be falling around averages rather than extreme. The findings thus reflect upon an expression of discontent that a public service provider may anyway face in the course of providing healthcare services. Therefore, by considering these responses as genuine expressions of satisfaction levels, the authors are not committing any bias. The analysis in the next section also confirms this conclusion.

To begin with, we are raising an important question: Do the respondents differentiate in the healthcare services based upon their characteristics? Therefore, the factors that are significantly associated with the satisfaction levels are first identified and then an attempt is made to understand how these important (significant) factors taken together explain the variations in the levels of satisfaction. Tables 5 through 7 provide evidences (chi-square and rank correlation) for significant variation in the relationship between satisfaction level(s) and the factors: (i) behaviour of health professionals, (ii) infrastructure and (iii) services and the following key characteristics of the respondents namely: (i) patient type, (ii) health institutions accessed, (iii) gender, (iv) caste, (v) income of respondent, (vi) distance travelled to access health institutions and (vii) cost incurred in availing services.

Table 7 provides evidence on association between patient satisfaction to health services of institution and their characteristic. Findings reveal that except for gender, other characteristics of respondents are significantly associated with the level of satisfaction services provided by health institutions: While the respondents are relatively dissatisfied with services provided by the first access point of health institutions, that is, PHC, SHC and CHC, they are comparatively more satisfied with the services provided by the government and private hospitals. The findings also reveal that respondents are satisfied with services provided by both government as well as private hospitals. Respondents also reported that in-patients services are more satisfactory than the services provided to out-patients. Scheduled Caste respondents are relatively less satisfied than other caste groups, while Scheduled Tribes are the least satisfied. Respondents in higher income category are relatively dissatisfied with the health services provided by institutions compared to the lower income respondents. Like behaviour of health professionals and quality of infrastructure of the health institutions, responses on satisfaction with services provided do not differ with gender or distance travelled.

To sum up, (i) the behaviour of healthcare professionals is significantly associated with health institutions, patient type, income, distance travelled and caste while gender has no association with behaviour of healthcare professionals; (ii) except gender, patient characteristics—institution type, income, caste patient category—are significantly associated with their satisfaction with infrastructure facilities available at the health institutions; (iii) except for gender, other characteristics of respondents are significantly associated with the level of satisfaction services provided by health institutions. Taking these characteristics of the respondents together, we try to understand how these interactions explain the variations in the satisfactions with the (a) behaviour of health personnel; (b) infrastructure and (c) services provided by health institutions. In doing this, we have also included, in the analysis, if the respondent has availed, from the institution, benefit of (i) medical tests (ii) supply of medicine and (iii) other health services. It should be noted that when two factors significantly interact in explaining the variations in satisfaction level, then neither of the individual factors, despite their being significant, lose their importance. In this situation, the only interpretation that makes sense is the simultaneous change of the two factors.

Table 8 gives the findings of ANOVA, which explains variations in satisfaction for the three independent variables: (i) behaviour of health personnel; (ii) infrastructure; and (iii) services provided by health institutions. After having various combinations of factors, the explanation emerging from Table 8 captures the most suitable and significant combination of explanatory factors. Other combinations do not make any significant impact on explaining variance in the satisfaction levels. Therefore, they are not part of the explanation. Moreover, if any factor is found significant in interaction with other factors, its independent relevance loses. Hence the explanations in Table 8 have not made use of individual explanatory factor(s).

Two significant interactions in explaining the satisfaction with services (Table 8, column 3) were (i) patient type and caste and (ii) cost incurred in accessing services and institutions. An attempt is made in this sub-section to understand how consumers' perceptions regarding the behaviour of health personnel and infrastructure influence these findings. Table 9 presents results of this three-way interaction.

Inclusion of perceptions of consumer in the explanation improves the R bar square substantially to about 38 per cent. This means that a substantial part of variations in satisfaction with services are explained by interactions among (i) patient type and caste; and (ii) cost incurred in accessing services and institutions along and perception on (a) behaviour of health personnel, and (b) infrastructure. What these three-way interactions tell is that:

These findings do indicate that perceptions on services provided are also shaped by, apart from other factors, perceptions about behaviour as well as infrastructure.

The analysis of the exit survey data from Madhya Pradesh suggests that client satisfaction responses in user exit surveys are not positively biased and are substantially critical. The criticality of responses is reflected significantly in the case of satisfaction on infrastructure and also on services provided by the healthcare institutions; evidences show that there is no positive bias for satisfaction on infrastructure as well as quality of services provided. However, the questions that seem to be answered with the least bias in the exit survey are those that are the least useful. If consumer perceptions in exit surveys closely reflect observable infrastructural facilities, questions on these perceptions would not add to what could be learned from direct observation using facility surveys.

What distinctively emerge from this exit survey are the subjective but positive perceptions on the behaviour of practitioners and the respondents' overall dissatisfaction with service quality. The perception is that public healthcare institutions at the sub-district level have poor infrastructure and poor service delivery and their health personnel display a dissatisfactory behaviour towards patients compared to the district level institutions, both public and private. This obviously calls for changing the functioning of SHC, PHC and CHCs both in terms of staff capabilities as well as service delivery. These aspects aside, the exit survey data are informative. Cost incurred to access services from the institution has a significant negative impact on the perceived value of the service. Another finding is that the characteristics of the patient seem to matter a great deal for reported satisfaction. They are able to differentiate satisfied from less satisfied healthcare consumers and shed light on the service quality that increase user satisfaction. It is highly disturbing that class, caste and income of the consumers should influence their perceptions about services of health institutions.

Importantly, ANOVA results conclusively show that responses are affected by interactions of factors rather than by individual factors. This analysis highlights three important interactions:

Three-way interactions also prove that perceptions on services provided are shaped by, apart from other factors, perceptions about behaviour as well as infrastructure.

Lastly, data in the exit surveys may also include (i) availability of essential drugs and their supply, (ii) quality of tests and the (iii) process of examining the patients. This is because the perceptions about services and behaviour of health personnel are also shaped by the following factors: Were the essential medicines made available on time? Who examined the patients? Was a physical examination done? Were the causes of ailment, general health and nutrition and instructions on medications discussed with patients? Did the diagnosis and healthcare lead to cure? These are objective assessments made by patients and may on the one hand become policy questions and may lead to objectivity in investigation, on the other.

The findings of this study may be used in two different ways. First, they could be used for joint action at the district and sub-district levels for identifying ways to strengthen the healthcare institutions. Second, they may be used as evidences for legal actions.

Accessibility of health services and its quality are key concerns in remote rural areas, especially for the poor and marginalized section of society. Owing to poor infrastructure, indifferent behaviour of health personnel, and poor quality of services in public institutions, even clients from poor income group are approaching the private institutions. Lack of medicine, erratic supply of electricity, non-availability of doctor, shortage of skilled health personnel, general apathy towards the poor, indifferent attitude of the health staff, poor maintenance of critical equipment are some of the factors that have deterred the patients, from all caste and income groups, from accessing the public health services both in urban and rural setting. The present standard of health services do not match the norms and hence when it comes to the question of quality service, people increasingly depend on the private healthcare providers.

As proponents of quantitative research, we are treating these responses not as given but shaped by the failure of the health institutions to service the needy. The findings of this research could be seen as the centre of deep subjectification at one level and voices of disapproval, on the other. The interaction of community with health institution, while seeking healthcare, takes shape not as a conflict between the providers and the sick; it stems from the question of governance. This interaction has less to do with the structure of the institution; it is more about the way in which the conduct of sick, as individuals, are predetermined. In doing so, institutions not only cover the legitimately constituted forms of subjection but also modes of actions—considered and calculated—which were intended to shape possible field of actions of the sick. In other words, to govern is to construct the possible field of actions of the sick. These disciplinary techniques, for controlling and internalizing the behaviour, also lead to self-discipline in the production of the individual as a subject (Gorden, 1980). This is not so much to exert power, but to ensure that the disciplinary mechanism is in place. Sovereignty and disciplinary mechanism are two absolutely integral constituents of the general mechanism of power (Gorden, 1980). Inevitably, this also represents asymmetrical power relations between healthcare providers and healthcare seekers.

Following Foucault (1980: 95–108), it can be argued that the privileged place to observe the power in action is the relations between the institution and the individuals (Balan, 2010)—How institutions exert their power on individuals and how the latter affirm their own identity and resistance to the effects of power because power is coextensive with resistance, producing positive effects. Consequently, despite strictly adhering to norms set by health institution in course of their interactions with institutions, the consumers also resist the way they are treated. Healthcare services become political for they represent the way we are governed. No matter how shabbily recorded, the findings of the client exit survey may be seen as consumers' protest to services provided by public health providers. These responses provide opportunity to understand the characteristics of liberalism and the mentality of administration that serve a society, existing external to the health institutions.

Information from healthcare consumers about the causes of their dissatisfaction with public health providers can become an important tool for monitoring quality of services, provided the method of data collection addresses the issues of validity of responses. Such Exit Surveys can help in protecting interests of the consumers only if their feedback affects the decisions of the health administration (Crow et al., 2002). District health plan may include findings of exit survey conducted, on a regular basis, either by an independent agency or by the local health department. Results of these surveys may be used as performance indicators at the state or national level in identifying consumers' perceptions and concerns, areas of service failure, and scope for improvement.