Users' perspectives on non-clinical quality of care in public and private primary healthcare in Albania

Background: Aiming to tackle the rise of non-communicable diseases and an ageing population, Albania is engaged in boosting primary healthcare services and quality of care. The patients’ perspectives on their experience with public and private providers are, however, missing, although their viewpoints are critical while shaping the developing services. Consequently, we analyse perceptions of users of primary healthcare as it relates to non-clinical quality of care and the association to sociodemographic characteristics of patients and the type of provider. Methods: A facility-based survey was conducted in 2018 using the World Health Organization responsiveness questionnaire which is based on a 4-point scale along eight non-clinical domains of quality of care. The data of 954 patients were analysed through descriptive statistics and linear mixed regression models. Results: Similar mean values were reported on total scale of the quality of care for private and public providers, also after sociodemographic adjustments. The highest mean score was reported for the domain ‘communication’ (3.75) followed by ‘dignity’ (3.65), while the lowest mean scores were given for ‘choice’ (2.89) and ‘prompt attention’ (3.00). Urban governmental PHC services were rated signicantly better than private outpatient clinics in ‘coordination of care’ (2.90 vs 2.12, p < 0.001). In contrast, private outpatient clinics were judged signicantly better than urban PHC clinics in ‘condentiality’ (3.77 vs 3.38, p = 0.04) and ‘quality of basic amenities’ (3.70 vs 3.02, p < 0.001). For the other domains, no statistically signicant differences were observed. ‘Autonomy’ was reported as least important attribute of quality. Enrolment in health insurance was a predictor of higher quality ratings (coecient = 0.06, p = 0.02). Conclusion: While the perception of non-clinical care quality was found to be high and similar for public and private providers, promptness and coordination of care require attention to meet patient’s expectations on good quality of care. There is a need to shift from a ‘paternalistic’ model to a ‘co-managing the illness’, raising awareness on autonomy.


Introduction
The goal of primary healthcare (PHC) is to provide universally accessible rst level services for individuals, families and communities thereby ensuring their referral to hospital and other specialized services when needed. It is increasingly acknowledged that the quality of services matters to progress towards Universal Health Coverage (UHC). 1,2 Patients 'viewpoints, perceptions and experiences, including non-therapeutic dimensions of care such as communication, attention, treatment or con dentiality, is a central aspect of quality of care. 1,2, 3,4 Albania, a south-eastern European country with a health system in transition, has been engaged for several years in improving PHC services to better address health system challenges, such as the steady rise of non-communicable diseases (NCDs). 5,,6 ,7 Indeed, the most recent Demographic and Health Surveys (ADHS) conducted in 2009 respectively 2018, indicate for example marked increases in the prevalence rates of hypertension. 8 In 2015 as part of its commitment to move towards UHC, the Ministry of Health and Social Protection (MoHSP) has introduced a free check-up programme offered for all citizens aged 35 to 75 yearsindependent of their insurance coverage or health condition . 5,9 In parallel, a mandatory health insurance scheme is in place, as part of the social protection system, covering medical examinations at the public and, more recently, at some private contracted health institution. Based on the results of ADHS 2018, around 37% of the population aged 15-59 years bene ts from entitlements through the mandatory health insurance scheme. 8 Based on the referral system and the drug reimbursement structure (i.e. full or partial) the scheme is free or subsidized with co-payment. 10 Public PHC is currently being provided in Albania through a well-established network of 413 urban and rural health facilities and additional health posts. A package of basic medical services such as (i) emergency care; (ii) health services for children; (iii) women of reproductive age; (iv) adults and (v) elderly people; (vi) mental health care; and (vii) health promotion and education is offered mainly through health centers. 10 In addition, there are private healthcare services. They may be for-pro t or not-for pro t providers. The for-pro t outpatient clinics (providing PHC and specialty care services), which are sometimes located in private hospitals, have experienced substantial growth in the last decade, especially in urban areas. In 2019, ten private hospitals, 229 specialized private diagnostics and laboratory clinics and 177 outpatients' medical centres and cabinets were licensed . 11 While adding a provider option for patients, they have proven challenging to the coordination of the national health system as, to some degree, they are duplicating both public PHC services and specialty services provided in polyclinics. 12 Governmental PHC has often been challenged by ill-equipped facilities, bypass to secondary or tertiary care, lack of trust in PHC personnel and access barriers to services (e.g. long waiting times and distance to PHC centres). 9,13,14,12 At a time of rapid change in health demands and growth in PHC providers it would seem important to understand what service users would value in terms of quality.
However, information on patient encounters' with PHC both in public and private outpatient settings and their perspective as it relates to the non-clinical quality of care is missing in Albania.

Quality of care and patient experience across settings
Quality of care is a broad concept which includes structural, technical, process and outcomes aspects. 15,16 Different instruments have been developed to measure and assess quality from different perspectives. A tool developed and validated for measuring and analysing the non-clinical aspects of care is the "health system responsiveness tool". 17,18,19 In 2018 World Health Assembly proposed a set of indicators and a framework for measurements in patient centeredness areas to aid countries in embedding patient experience as an assessable and reportable component of quality. 20 Both the public and private health organization are engaged in improving quality of care. The private sector is often viewed as more client-centered, 21 better at patient education and interpersonal satisfaction, 22 and patients seem to experience better timeliness and hospitality. 23 Quality of care is comparatively well rated in public health services in some western European countries. 24,25 Other studies nd it di cult to draw clear deductions about the advantage of any particular type of settings, instead associating quality of care with the main features of health service provider such as the organization of and remuneration for services, number of skilled health specialists rather than the clear ownership of the health facilities. 26,27 Recent systematic reviews led to contradictory conclusions in respect to quality of care differences between public and private providers due to different review methodologies and, above all, diverse settings and contexts where private providers played different roles within the health system. 28 Several research studies have shown that rural populations are more reluctant to express discontent and are generally more satis ed with quality of care and health-care systems than their urban counterparts, 29 though rural patients appear to mirror an undervaluing of primary care in favour of specialty care. 30 Primary health care in many rural areas have the challenges of; staff recruitment, poor physical amenities; lack of accessibility to diagnostic health services which undermines quality and effective care when contrasted to urban settings. 31,32 Additional evidence is required with regard to public and private users 'experiences with quality of care and operating providers' characteristics, 33 factors that drive quality variations and quality improvement approaches 20 in order to guide PHC strengthening. To date, substantial published work is available in this eld 34,35,36, 37, 38 but none in Albania to the best of our knowledge.
In the light of the commitment of the Albanian Government to improve UHC, this study analyses perceptions of the users of public and private PHC services pertaining to non-clinical quality of care and their associations with the sociodemographic characteristics of patients and the type of healthcare provider.

Methods
Study setting, design and sampling All governmental health centres (HC) in rural and urban areas in two regions of Albania, Fier and Diber, were included in the study. The study aimed to include all private health providers offering outpatient / ambulatory services in the regions covered. In one region (Diber) there were no private services and in the other were 8 (Fier). From these 8 private clinics 5 ve consented to participate. Therefore, 38 public healthcare facilities and 5 private healthcare facilities were included in the sampling. The sampling for the two regions was done based on the number of visits per HC in 2017 received from regional insurance directorates.

Questionnaire on patient perception
We assessed patients' perceptions on eight domains of responsiveness and service quality through the World Health Organization (WHO)'s Health System Responsiveness Questionnaire, a publicly and freely available tool 39 which has been widely used in various settings 17,34,36,40,37,41 . The tool is structured along eight domains: (i) autonomy; (ii) choice of health care provider; (iii) clear communication; (iv) con dentiality; (v) dignity; (vi) prompt attention; (vii) quality of basic amenities; and (viii) access to social support networks. We excluded the domain of 'access to social support networks' because our study focused on users of outpatient services. Based on our critique and as suggested by other researchers in this eld, we added 'coordination of care' as an additional domain, given the importance of patients with chronic conditions in the Albania setting. 38 All individual items were scored on a scale from 1-4 (1 = bad; 2 = rather bad; 3 = good; 4 = very good).
Patients were also asked to choose the domain they consider as 'most important' when consulting a healthcare provider. Hence, in this study, we make a distinction between two categories of users' measures of non-clinical quality of care: (i) patients' most recent experience (the level of responsiveness as measured by the interactions that patients have with the healthcare provider) and (ii) patients' expectations on attributes of quality (patients' evaluations of what is considered important when receiving care in general, relative to their expectations).
The WHO questionnaire was translated from English to Albanian and then translated back to English prior to conducting the interviews. We changed the word 'con dentiality' and translated as 'privacy respected' and 'autonomy' as 'involvement in health decisions' in order to be easy comprehended and to be closer to laymen comprehension of the terms. Also, on 'coordination of care', we slightly changed the item 'the physician knows if certain tests have to be conducted regularly' into 'doctor knows your medical history (main developments on illness) and 'you were helped (feel assisted) to transit from one provider to the other'. Patients' socio-demographic characteristics such as age, gender, employment status, education, status of health insurance coverage was also collected in addition to the patients' health conditions including self-reported health status (poor/not poor) and the presence of chronic health condition (yes/no). In order to determine the internal consistency of the questionnaire, Cronbach's alpha was calculated.

Data collection
Data collection took place from July to August 2018. For study inclusion, participants had to be at least 18 years old, and they had to have had some form of outpatient care on the day of the interview.
Interviews were conducted by medical students that had completed at least a bachelor's degree. Prior to data collection, interviewers were trained for three days and the questionnaires were pilot tested in a different population but in the same regions. Data collection was done electronically using tablets through Open Data Kit (ODK) platform. Participants' responses were uploaded in a secured server at Swiss Tropical and Public Health Institute (Swiss TPH), Basel, Switzerland on the same day of the data collection and regular data quality check was conducted.

Data analysis
In a rst step, the characteristics of patients were compared across the three types of health facilities: (1) public urban PHC clinics, (2) private outpatient clinics (3) public rural PHC clinics. Mean scores of each domain, representing patients 'experience with quality were obtained using the margins syntax of Stata. Our primary analysis focused on the association between perceived non-clinical quality of care and type of facility. Factors associated with patient's perceived quality of care were included as potential confounders of this association. They were: (i) age; (ii) gender; (iii) education; (iv) occupation; (v) insurance status (yes/no); (vi) self-rated health (poor/good); (vii) presence of one or more chronic condition(s); and (viii) utilization of clinics over the past three months. Linear mixed models, with random intercepts for districts and facilities nested in districts, were thus used to investigate the association between the utilization of the type of health facility and non-clinical quality of care domains adjusting for patients' sociodemographic characteristics. The score of overall quality was determined as the mean of all available sub-scores requiring that at least six of the eight sub-scores were present. Analyses were repeated in a subsample of patients who had no missing sub-scores and the respective results showed only minor differences. The statistical signi cance level used in this study was 0.05. Data was analysed using Stata Statistical Software, version 15.

Results
Patients' characteristics by type of healthcare provider Out of 1083 were eligible study participants 954, accepted to be interviewed corresponding to a response rate of 88%. The characteristics of the study participants are shown in Table 1. There was a slightly higher proportion of female visits compared to male visits (58% vs 42%). The mean age of participants was 37 ± 20.1. Most of the patients were in the age group 18-60 years (54%). About 33% of the participants were unemployed and 50% had basic education (primary and secondary school, 5 to 9 years of study).
Almost 60% of the all respondents stated that they suffered from at least one chronic health condition.
The percentage of patients who reported suffering from two or more chronic conditions was higher in public rural PHC clinics (21%) and public urban PHC clinics (19%) compared to private clinics (8%). The proportion of patients recently diagnosed with a chronic condition (i.e. less than two years ago) was higher in private clinics (31%) compared with the governmental PHC facilities (vs 19%) (Table 1). When ranking the importance of domains for consulting by type of provider, among all patients included in the survey by total, 'communication' was rated highest, followed by 'dignity' (Figure1). The domains seen as most important, communication and dignity, also received the highest mean quality ratings. The 'prompt attention' domain was considered as an important domain by patients (ranked third); at the same time, it was perceived as poor and received low mean scores. 'Autonomy' and 'choice of provider' were least frequently reported as most important domain 'autonomy' also demonstrating lower mean ratings compared to other domains. Predictors of non-clinical quality of care Table 3 presents the results of multivariable analyses of the eight domains quality scores after adjustment for socioeconomic and health conditions of respondents. The mean total score showed no statistically signi cant difference between patients by type of health care provider (coe cient = 0.12, p = 0.27). However, the adjusted mean total score was signi cantly higher among rural patients compared to those consulting a public urban HC (coe cient = 0.2, p = 0.01).
Patients consulting a rural PHC manifested a higher average rating on the domains of 'coordination of care' (coe cient = 0.5, p < 0.01) and 'choice' (coe cient = 0.5, p = 0.01) compared with patients from public urban PHC facilities.
The mixed model results showed that patients attending private providers reported lower scores on coordination of care (coe cient = -0.62, p < 0.01) and prompt attention (coe cient = -0.24, p = 0.03) in comparison to urban PHC clinics but reported a higher average scores of con dentiality (coe cient = 0.45, p = 0.03), quality of basic amenities (coe cient = 0.81, p = 0.002) and autonomy (coe cient = 0.48, p = 0.07). Possession of a health insurance card was associated with a higher mean total score on quality of care (coe cient = 0.06, p = 0.02). Also, more frequent utilization of the facility was associated with higher mean values of the domains of 'coordination of care' (coe cient = 0.12, p = 0.01) and -by trend -on autonomy (coe cient = 0.12, p = 0.07).
Patients with one or more chronic health conditions reported a lower average score on the 'prompt attention' domain compared with those not suffering from any NCD (coe cient = -0.16, p = 0.03). However, they perceived experiencing better coordination of care when compared with healthy participants (coe cient = 0.16, p = 0.001; coe cient = 0.2, p = 0.001).
Possession of a health insurance card was associated with a higher mean total score on quality of care (coe cient = 0.06, p = 0.02). Also, more frequent utilization of the facility was associated with higher mean values of the domains of 'coordination of care' (coe cient = 0.12, p = 0.01) and -by trend -on autonomy (coe cient = 0.12, p = 0.07).
Patients with one or more chronic health conditions reported a lower average score on the 'prompt attention' domain compared with those not suffering from any NCD (coe cient = -0.16, p = 0.03).

Discussion
This study has offered insight into the non-clinical quality of care attributes, simultaneously by measuring patients 'experience and expectations of quality among public and private PHC providers respectively, in two regions of Albania. The ndings of this study are discussed based on the relevance of the main non-clinical quality attributes (domains) and their perceived responsiveness by healthcare provider type.
Quality of care by type of health care provider The study indicates that urban public PHC services and private outpatient clinics do perform similarly in respect to attributes of non-clinical quality of care. This is in contrast to, Bleich and colleagues who identi ed lower quality rates amongst private healthcare users compared to public counterparts, in a study conducted in 21 European Union countries . 25 These results deviate from ndings of other studies indicating that private health facilities appear to be of higher (interpersonal) process quality, including responsiveness and effort, and conceivably being more patient-orientated than public facilities. 21,22,23 Our results do suggest that while the overall quality ratings were similar, private providers are rated better, on quality of basic amenities, con dentiality and autonomy. Although con dentiality was well rated in terms of responsiveness (patients' experience when receiving care, as assessed by the mean sores), (Table 2, Figure 1), this domain was considered by respondents as of comparatively low importance for consulting a PHC service (Figure 1). This aligns to other studies which identi ed con dentiality being a neglected aspect of care in less developed countries, partly due to lower importance given to this domain, mainly attributed to resource limitations and lack of awareness. 42 One reason why con dentiality is better perceived in private sector in the present study could be linked to relatively good infrastructural The attitudes of the rural patients interviewed within the present study, may also mirror the past communism area positions of the society with strong community ties, where traditionally state-owned institutions, doctors included, were the ultimate and unquestionable authority. The ndings may also re ect a lack of awareness of patients' rights and weak, undeveloped patients' centeredness concepts.
Nevertheless, when compared with urban patients in public facilities, rural patients gave considerably lower mean values on prompt attention (Table 2), perceiving di culties related to geographic proximity, means and costs of transport and waiting time. However, after adjustments for socioeconomic and health characteristics, these results did not yield a statistical signi cance when compared to urban public providers (Table 3).
Patient-provider interaction; the path toward informative and participatory decision-making. when encountering the health providers and health system. Also, in terms of importance of domains, as depicted from patients' theoretical point of view, our study showed that patients, independent from using was frequently rated as one of the least important domains (Figure 1). This shows that involvement of patients in treatment choices is still an evolving area, especially in the governmental sector. Some efforts are being invested by private outpatient clinics in giving patients more information about alternative types of treatments and tests; however, in our study there were not any noteworthy differences compared to the public sector. It has been argued that low attention to autonomy can be explained by persistent paternalistic behaviours of both provider and patients regarding their position with each other and within the health system. 47 Patients' voice in healthcare delivery process and community involvement on quality of care improvement initiatives are latent in Albania and have yet to be actively developed in the health system to support health policymaking. 5 Prompt attention, choice and coordination of care When averaging ratings of responsiveness for the total population, domains of 'ability to choose doctor', 'prompt attention' and 'coordination of care' were among the lowest rated, suggesting a poor performance. 'Prompt attention' was ranked third by the level of importance and it was the second lowest rating by the level of responsiveness (Figure 1). This indicates that this domain is of high importance to patients but not experienced satisfactorily in the frame of their most recent PHC consultation, especially in rural settings.
Respondents receiving care from private outpatient clinics provided a low average score on the 'prompt attention' domain. This may be related to two facts; rst, private outpatient clinics operate during some hours of the day, by inviting several specialist doctors, sometimes from the capital city of Tirana or nearby countries such as doctors from Greece, Italy, or Turkey. Second, the patients attending private clinics frequently do not bene t from health insurance coverage. They may be living in close or distant villages, and they may have gone through different processes and obstacles within the public health care system prior to ending up at the private clinics. Moreover, doctors working at the same time at the public and private health sector (dual practices) and patient juggling are a concerning phenomenon, especially in low-and middle-income countries, 48 making private PHC services not constantly available due to limited presences of doctors, hindering thus prompt attention, choice and good coordination of care.
The factors mentioned above might also explain the discontent of patients with their ability to choose the provider they want to consult about their medical condition even when choosing or transiting to the private health providers.

Sociodemographic characteristics and perceived quality of care
Our results showed that being in the possession of a health insurance card, having utilized the health service more recently and being a pensioner, were good predictors of positive quality ratings of the domains of 'coordination of care' and 'involvement in healing options' (autonomy). At the same, suffering from one or more chronic conditions were good predictors of lower quality ratings on the 'prompt attention' domain and higher ratings for coordination of care. Thus, while shaping the new service delivery models, policy makers and public health researchers should emphasis the ways in which to deploy health workers and how to engage patients in treatment choices in order to deliver wellcoordinated care. As Albania moves towards a better coordinated health service it will be of importance to build up an electronic medical record system so that different providers can share and exchange relevant patient information.
In the current study, people in good health were overall more critical of the quality of care, especially with the ability to choose a provider, while gender and age were no predictors of quality of care ratings.
Previous studies have yielded mixed results between perceived quality and patients' sociodemographic and health characteristics. 25,34,36,47,46,49 Low perceived quality of care has been associated with users in poor health, uninsured people or users that have made fewer visits to providers. 25 Other studies associate perceived good quality with older age and higher income, 37 self-reported good health status and rural residency. 29 Patients' perceived quality differences may also be explained by variability of the quality instruments and the context on which the instruments have been implemented, factors related to individual characteristics, previous encounters with health care providers and the cultural, historical and geographical environment.

Study limitations
This study has some limitations. The number of respondents who answered each item, varied from one domain to the other (selecting for example the 'non applicable' answer option), re ecting variable understanding and sensitivity of respondents to eight domains of WHO responsiveness tool. Therefore, when data was analysed, a condition was set to have at least 75% of the all domains (six out of eight) entirely answered. However, the mean values trend did not change even when we ran the analysis under different domains ful lled rather than six. The study was conducted applying a widely accepted and validated tool, however, entirely relying on self-reporting perceptions rather than measurements of an impartial observer; therefore, possible variations of the patients' perceived non-clinical quality may be attributed to differences in patients' characteristics, cultural aspects, previous experience and expectations rather than actual provider practice. The response rate was quite high (88%), but certain categories of patients, for example younger patients did not give consent to participate. Therefore, a certain response bias cannot be excluded. Further, while all governmental providers in the two regions covered by the study participated, three out of eight private out-patient provides did not consent to participate in the study.
The two regions covered by the study make up around 16% of the territory of Albania and demographically represent around 15.7% of the population. One region (Diber) represents the mountains relatively poor part of the country while the second (Fier) is characteristic for the coastal, partially industrial settings in the country. Thus, the two regions re ect two patterns of Albania but cannot be considered as fully representing the socio-cultural and economic diversity of the country. Consequently, we cannot exclude the possibility that if applied to other regions, namely the urban context of the capital city Tirana, the results would differ.

Conclusion
The overall perception and ratings of non-clinical quality of care by patients is similar across those consulting governmental and private services respectively in urban and rural areas. Respondents rated their experience with quality of care high, indicating that this dimension matters for consultation. Other dimensions of importance for patients independently of the type of service provider were communication and dignity. Opposite, users of public rural PHC services rated their experience with "prompt attention" considerably lower than those consulting public urban PHC services. Patients attending private outpatient clinics rated the dimension 'coordination of care' as low.
Patients' involvement in their care was not prioritised by patients and considered as a less important aspect of quality. This would suggest that relationship between health care provider and the patient should change from a 'paternalistic' model to a 'co-managing the illness' model, where both patients and caregivers meaningfully participate in decisions related to the healing process. Educating patients and doctors to make the most of their interactions would be an effective way to tackle the low awareness on autonomy.
Given the ageing population and concomitant rise of chronic health conditions, home care models, might be bene cial to reach out into communities and raise promptness of response of services. In the absence of a consolidated health information system, lack of timely exchange of patients' medical records, within and between public and private healthcare providers, members of the allied health workforce such as