|
Description
| Project Summary
Introduction: Responsiveness of human resources for health (HRH) is defined as the social actions that health providers do to meet the legitimate expectations of service seekers. It may dissuade patients from early care seeking, diminish their interest in adopting preventive health information, decrease trust with health service providers, and marginalize at-risk population groups. The overall goal of this research was to examine HRH responsiveness in rural Bangladesh, to develop a scale to measure the responsiveness, and finally to demonstrate the application of the measurement method.
Methods: Data collection took place in rural parts of Khulna, a southwestern division of Bangladesh. The qualitative portion consisted of in-depth interviews (IDI) with physicians (seven public, five private, five informal), in-depth interviews with clients (n=7), focus group discussions (FGD) with clients (two sessions each with males and females), and observations in consultation rooms of public, private, and informal sector healthcare providers (one week in each setting). The quantitative research consisted of structured observation of 393 physicians (195 from public and 198 from private sector). This data was collected for developing a scale of responsiveness through exploratory factor analysis (EFA), involving 64 items (generated through the qualitative part of this project). Inter-rater reliability was assessed by same three raters observing 30 consultations, using the scale (later named as Responsiveness of Physicians Scale or in short ROP-Scale). Study data were collected between August 2014 and January 2015. Qualitative data were analyzed by the framework analysis method. World Health Organization’s (WHO) health systems responsiveness framework was modified, based on literature review and expert opinions, to include the following domains for qualitative analysis: Friendliness, Respecting, Informing and guiding, Gaining trust, and Optimizing benefit. Quantitative data were analyzed by EFA, followed by assessment of internal consistency by ordinal alpha coefficient and inter-rater reliability by intra-class correlation coefficient (ICC). For comparing responsiveness of public and private sector physicians two sample t-test, multiple linear regression (MLR), multivariate analysis of variance (MANOVA), and descriptive discriminant analysis (DDA) were used.
Results: User and provider perceptions of responsiveness of physicians in rural Bangladesh often overlapped but at times diverged. Due to high patient load, physicians in the public sector usually failed to spend enough time with patients for proper history taking, asking questions, examining, and reassuring. Although not satisfactory, according to patients in qualitative part of the research, physicians in the private sector were more responsive towards the patients, especially in terms of conducting examinations with care, asking questions, and giving little reassurance. Most of the patients complained that physicians in general (i.e., both in public and private sectors) were not responsive, especially in terms of talking to them enough, compassionately touching them (for examining, for giving reassurance), and explaining their condition. They also complained of losing trust in physicians, as they seemed not to be caring, but businesslike. Patients demanded that, in order to be responsive, physicians should not only be prescribing drugs, but also be sensitive to patient’s financial status. Physicians should tell them the cost of treatment, try to understand whether patients can afford it, and if necessary, tailor the treatment accordingly. On the other hand, physicians also acknowledged their inadequacies, but attributed these to the overall health systems constraints, patient loads, lack of proper training on responsiveness issues, and often abuse by the patients. Psychometric analyses identified 34 items grouped under five domains (or subscales) to constitute the Responsiveness of Physicians Scale or, in short, ROP-Scale. The five domains, derived through EFA and later named through discussing with the relevant experts, are as follows: Friendliness, Respecting, Informing and guiding, Gaining trust, and Financial sensitivity. The scale has a very high internal consistency with ordinal alpha coefficient of 0.91. Inter-rater reliability was also very high with intra-class correlation coefficient (ICC) (2, 3) of 0.84. The scale also demonstrated face validity (through expert consultation), content validity (through qualitative research and literature review) and criterion validity (concurrent validity by correlation coefficient of 0.51 with consultation time; and known-group validity by comparing public and private sector physicians’ responsiveness with private sector scoring 0.18 higher mean score). The study found the mean responsiveness score of public sector physicians to be1.98 and that of private sector physicians 2.16; and the difference statistically significant in t-test with t statistic of -6.04 (p-value <0.01). The difference remained statistically significant in the multivariable models after adjusting for the confounding covariates such as age, gender and local origin of the physician and age, gender and level of education of the patient. Qualitative data added value to this finding by suggesting that, despite slightly better responsiveness of private sector physicians, none of the sectors were sufficiently responsive, according to service seekers. In domain-specific evaluation of responsiveness, the public sector outperformed the private sector in domains of Gaining trust and Financial sensitivity. The domain Respecting was identified in DDA as the most important domain in dividing the public and private sector based on responsiveness. The qualitative part of the study found the private sector physicians to be more tolerant, polite, and courteous than the public sector physicians, as opined by patients. Nevertheless, private sector physicians were criticized by patients for attending more patients than their capacity, prescribing more diagnostic tests, and showing reluctance to refer patients who they failed to treat. Qualitative findings supported the quantitative findings that public sector physicians were more prudent in gaining trust and being financially sensitive to the patients.
Conclusions: This study demonstrated the detailed process of development and application of a psychometrically validated ROP-Scale. In this process, I reviewed the earlier work on health systems as well as HRH responsiveness, defined the HRH responsiveness, discussed caveats in different aspects of understanding and measuring responsiveness, proposed a conceptual framework to examine HRH responsiveness, identified five domains of HRH responsiveness, presented the findings across the domains of responsiveness, and compared the responsiveness of public and private sector physicians’ responsiveness.
Data Abstract: The qualitative part of this research consisted of in-depth interviews (IDI) with physicians (seven public, five private, five informal), in-depth interviews with clients (n=7), focus group discussions (FGD) with clients (two sessions each with males and females), and observations in consultation rooms of public, private, and informal sector healthcare providers (one week in each setting). The quantitative research consisted of structured observation of 393 physicians (195 from public and 198 from private sector). First, I approached the respondents after preparing a list of all physicians working both in public and private sectors in the district. For IDIs with clients, I followed heterogeneous purposive sampling, with maximum variation in age, gender, level of education and occupation. I generated a list of potential respondents with inputs from local residents (personal contacts) and contacted them. Selection criteria for respondents were: >18 years age, consulted a physician at least twice in lifetime, with the last consultation within one year. For FGDs, I followed homogenous purposive sampling, attempting homogeneity in terms of gender (and also profession in case of females). Female FGD respondents were selected from the female employees of two local educational institutions (a school and a college). Selection of sites for observation was based on principles of convenience sampling (feasibility of travel at different times of the day) as well as purposive sampling (ensuring coverage of both public and private sectors). For quantitative data collection, a list of all physicians in Khulna district who were likely to be present during the data collection period was prepared beforehand. I chose the census method, as there were not sufficient physicians for sampling.
The qualitative data were collected in order to explore the perceptions and practices of outpatient healthcare users and providers (physicians) regarding the elements of responsiveness of physicians in rural Bangladesh. In order for triangulation, data were collected from both the service providers and the service seekers. From among the service providers, in-depth interviews were done with the following types of respondents: 1. In-depth interviews with seven public sector physicians 2. In-depth interviews with five private sector physicians 3. In-depth interviews with five informal providers (village doctors) From among the service seekers, both in-depth interviews and focus group discussions were done: 4. In-depth interviews with seven patients 5. Focus group discussions with four groups of patients (two sessions each with males and females) In addition, observations were conducted in the following manner: 7. Observation in the consultation settings of public sector physicians for seven days 8. Observation in the consultation settings of private sector physicians for seven days 9. Observation in the consultation settings of informal providers (village doctors) for seven days Quantitative data were collected through structured observation of 195 consultations in the public sector and 198 consultations in the private sector. These data were consulted as these persons (physicians from different sectors and patients) are the direct stakeholders in understanding responsiveness of physicians. Other potential respondents would be the policy makers, district health managers, medical teachers, etc. They were not consulted due to time and resource constraint, and also because, it was not essential to address the research questions. These data are being shared as we believe that sharing the research data would improve the transparency. This would also allow other researchers to take advantage of the data and conduct further analysis. Two Research Assistants (RA) helped in data collection and transcription. One of them was a male anthropology graduate, and the other was a female with same background. The male RA conducted four In-depth interviews, and the female RA conducted two focus group discussions with the female respondents; out of a total of 24 interviews and four FGD sessions. They prepared all the transcripts and Word processed the field notes from the observations of three weeks duration. In quantitative part of the research, 20 RAs collected data by structured observation method.
In-depth interviews and focus group discussions were audio recorded. These were all transcribed verbatim. Detailed notes were taken too. Field notes of observations were taken as well. Notes were taken during the interviews, FGDs, and observations. Observation field notes were elaborated later, within 24 hours of the observation. We are sharing only the qualitative data in this platform. We will share the quantitative data in other relevant repositories. Due to human participant constraints, only a subset of the qualitative data (interviews with providers) can be shared. Consent obtained from other participants precludes sharing of observations and interviews. |
