Patient satisfaction with HIV/AIDS care at private clinics in Dar es Salaam, Tanzania

Abstract

Health system responsiveness measures quality of care from the patient’s perspective, an important component of ensuring adherence to medication and care among HIV patients. We examined health system responsiveness in private clinics serving HIV patients in Dar es Salaam, Tanzania. We surveyed 640 patients 18 or older receiving care at one of 10 participating clinics, examining socioeconomic factors, HIV regimen, and self-reported experience with access and care at the clinic. Ordered logistic regression, adjusted for clustering of the clinic sites, was used to measure the relationships between age, gender, education, site size and overall quality of care rating, as well as between the different HSR domains and overall rating. Overall, patients reported high levels of satisfaction with care received. Confidentiality, communication, and respect were particularly highly rated, while timeliness received lower ratings despite relatively short wait times, perhaps indicating high expectations when receiving care at a private clinic. Respect, confidentiality, and promptness were significantly associated with overall rating of health care, while provider skills and communication were not significantly associated. Patients reported that quality of service and confidentiality, rather than convenience of location, were the most important factors in their choice of a clinic. Site size (patient volume) was also positively correlated with patient satisfaction. Our findings suggest that, in the setting of urban private-sector clinics, flexible clinics hours, prompt services, and efforts to improve respect, privacy and confidentiality may prove more helpful in increasing visit adherence than geographic accessibility. While a responsive health system is valuable in its own right, more work is needed to confirm that improvements in health system responsiveness in fact lead to improved adherence to care.

Introduction

Health system responsiveness (HSR) measures quality of care from the patient’s perspective, focusing on eight domains: access, confidentiality, promptness, dignity, autonomy, support systems, quality of basic amenities and communication (De Silva & Valentine, 2000). A number of these domains including communication, dignity, promptness, and access have been associated with adherence or follow-up in HIV patients. (Flickinger, Saha, Moore, & Beach, 2013; Geng et al., 2010; Miller, Ketlhapile, Rybasack-Smith, & Rosen, 2010; Poles et al., 2012; Schneider, Kaplan, Greenfield, Li, & Wilson, 2004). Previous studies have described patients’ ratings of HSR and factors associated with higher satisfaction in public-sector HIV clinics in resource-limited settings (Hirschhorn et al., 2010; Kagashe & Rwebangila, 2011; Peltzer, 2009; Wouters, Heunis, van Rensburg, & Meulemans, 2008). While studies have described technical quality of HIV care in the private sector (Innes et al., 2012), reports on HSR in private-sector HIV care are more limited.

Tanzania has the 14th-highest HIV prevalence in the world (UN AIDS, 2012). In Dar es Salaam, the largest city (population over 4.3 million), 6.9% of adults age 15–49 are HIV-positive (Measure DHS, 2012; Tanzania National Bureau of Statistics, 2013). In Tanzania, the private sector provides a substantial proportion of HIV care, accounting for 48% of HIV care expenditures (Sulzbach, De, & Wang, 2011). We report levels of patient-reported HSR at 10 private sector HIV care clinics in Dar es Salaam.

Methods

Population

The study was conducted in 10 private sector clinics receiving support from the US President’s Emergency Plan for AIDS Relief (PEPFAR)-funded Management and Development for Health. Patients over 18 years old receiving care at participating sites were eligible for inclusion. Sample size calculations were done based on our experience with a study in the public sector clinics with a maximum of 640 needed to estimate the ratings to +/−8% with a 95% confidence interval. (Poles) A convenience sample was enrolled at each clinic, with 40 patients per site in smaller clinics (< 10 patients/day), 65 in medium, and 75 in large clinics (> 20 patients/day).

Survey development and administration

The patient satisfaction survey used by the authors in a study of public-sector HIV clinics in Dar es Salaam (Poles, et al., 2012) was modified to reflect differences in service structure at private clinics. The survey was developed based on the World Health Organization (WHO) HSR framework (De Silva & Valentine, 2000) and review of relevant literature, and translated into Swahili. After pilot testing, the survey was finalized and back-translated for review (Poles, et al., 2012).

HSR domains and overall quality were rated on a 5 point Likert-scale ranging from poor to very good. Overall quality was also asked as a single question. The final survey was self-administered and captured patient demographics and self-reported care experience.

Patients were approached at the end of their visit by a research nurse who explained the study and obtained consent. The study ran from May to June 2011.

Data management and analysis

Data were entered twice into a database and compared to ensure data quality. All analysis was conducted in Stata 12 (Stata Corp., College Station, Texas, USA). Descriptive statistics were used to describe demographics and patient-reported HSR. Ordered logistic regression, adjusted for clustering of the clinic sites, was used to measure the relationships between age, gender, education, site size and overall quality of care rating, and between the different HSR domains and overall quality rating. P value <0.05 was considered significant.

The study was approved by the Institutional Review Boards at Harvard School of Public Health, Harvard Medical School, Muhimbili University of Health and Allied Sciences and the National Institute of Medical Research in Tanzania.

Results

Demographics

All patients approached completed the survey (n=640); 66% were female and mean age was 41 years (range 19–71; Table 1). Education levels were relatively high, with 39% having completed secondary school, and 20% with a university or professional diploma.

Table 1.

Demographics of respondents and results of exit survey of health system responsiveness

Measure	%	(n)	Range of clinic rates
Patient demographics1
Age
40 years or younger	51	(324)
Over 40 years	49	(316)
Female	66	(422)
Education
Primary	41	(257)
Secondary	39	(247)
Above secondary	20	(128)
Patient-reported health care experience
Travel Time to clinic
Less than 30 min	41	(264)
30–60 min	32	(202)
More than hour	27	(173)
Difficulty in getting to clinic
Moderately difficult, difficult, or very difficult	48	(303)
Minor difficulty	41	(258)
Not difficult	12	(73)
Time taken at facility
Less than 1 hr	75	(474)
1–2 hrs	20	(126)
Greater than 2 hours	5	(33)
Difficulty keeping appointments in past yr
Sometimes, often, or most of the time	11	(69)
Rarely	34	(212)
Never	55	(340)
Reasons for difficulty keeping appointments (top 4 reasons; answered only by those who missed an appointment)
Traveled out of town	38	(76)
Busy with work	22	(44)
Trouble getting permission from boss/husband/wife	12	(23)
No bus fare or transportation	10	(19)
Reasons for choosing this facility (top 4 reasons; patients could list multiple reasons)
Staff provides good service	59	(377)
Better confidentiality	28	(182)
Nearest to home or work	24	(151)
Better facilities than nearer one	17	(111)
Overall ratings of health care and selected health system responsiveness domains
Communication with health care provider "very good"	67	(435)	33–80%
Experience of being treated with respect "very good"	63	(405)	33–80%
Provider skill "very good"	63	(406)	35–88%
Promptness of attention at clinic "very good"	46	(296)	20–63%
Experience of confidentiality "very good"	70	(449)	28–93%
Overall health care rated "very good"	60	(382)	28–79%
Want to continue at this facility	100	(639)	100–100%
Would recommend this facility to others	94	(601)	86–100%

¹Due to missing data on some surveys, some totals do not add to 640.

HSR domains

Most responsiveness domains were rated highly, although considerable range between sites was seen: Seventy percent (site range 28–93%) of respondents rated their experience of confidentiality as “very good,” while 67% (site range 33–80%) rated communication with their health care provider as “very good”, and 63% (site range 33–80%) rated being treated with respect as “very good” (Table 1). Promptness received the lowest rating, with only 46% rating promptness of attention as “very good.” Reflecting these ratings, decreased waiting time was a frequent suggestion for how to improve the clinic, although 75% spent < 1 hour and only 5% > 2 hours at the clinic.

Although only 27% had to travel more than an hour to the clinic, almost one half (48%) reported moderate or significant difficulty getting to the clinic for their appointment. Almost one-half of respondents (45%) reported missing an appointment in the past year. Of those who had experienced difficulty keeping appointments, 38% reported travel away from home causing a missed appointment. Other common reasons included being busy at work, having trouble getting permission from a boss or spouse, and lacking transportation money. The most common suggestions for improvements were more space (28%) more staff (22%) and decreased waiting time (19%). Patients had a range of reasons for choosing their clinic, with 59% noting good service, 28% confidentiality, and 24% proximity to their home or work, and virtually all wanted to continue care at that site.

Factors associated with overall patient-rated quality of care

Ratings of overall quality of care were high, with 60% rating their overall care as very good (site range 28–79%). In a multivariate model, larger site size (medium: OR=3.02, 95% CI 1.35–6.75; large: OR=4.46, 95% CI 2.29–8.70) was associated with higher overall ratings of quality of care, while older patients tended to give lower ratings of care (OR=0.98 per year of age; 95% CI 0.962–0.995) (Table 2).

Table 2.

Multivariable model of association of patient factors and site size with overall quality of care (2a) and relative association of other domains of HSR with overall quality of care (2b)

Factors	Odds ratio	95% CI	P value
Association of patient factors and site size with overall quality of care
Female gender	0.74	0.41 – 1.35	0.283
Education [primary school only = reference population]
Secondary	1.16	0.73 – 1.84	0.49
Above secondary	1.62	0.93 – 2.80	0.079
Age [in years]	0.98	0.962 – 0.995	0.017
Site size [small = reference population]
Medium	3.02	1.35 – 6.75	0.012
Large	4.46	2.29 – 8.70	0.001
Relative association of other domains of HSR with overall quality of care
Respect rated "very good"	3.00	1.15 – 7.82	0.029
Confidentiality rated "very good"	2.74	1.37 – 5.48	0.010
Promptness rated "very good"	2.08	1.13 – 3.88	0.025
Communication rated "very good"	1.97	0.87 –4.48	0.094
Provider skill rated "very good"	1.40	0.89 – 2.20	0.127
Female gender	0.62	0.25 – 1.56	0.272
Education [primary school only = reference population]
Secondary	1.02	0.58 – 1.77	0.945
Above secondary	1.56	0.70 – 3.49	0.244
Age	0.98	0.95 – 1.01	0.173
Site size [small = reference population]
Medium	3.09	1.59 – 6.01	0.004
Large	3.92	1.90 – 8.06	0.002

Adjusting for patient demographics and site size, respect, confidentiality, and promptness were significantly associated with overall rating of health care, while provider skills and communication were not significantly associated (Table 2).

Discussion

We found that patients receiving HIV care from private-sector clinics in Dar es Salaam were generally very satisfied with care. HSR domains particularly highly rated included confidentiality, communication and respect. As opposed to other HSR studies (Hirschhorn, et al., 2010; Peltzer & Phaswana-Mafuya, 2012), older age was associated with some decrease in overall ratings of care. Possible explanations may include unmeasured differences in older individuals at private versus those at public clinics in our earlier study (Hirschhorn, et al., 2010) and overall a focus on a younger population in our study compared with Peltzer. Other results were not substantially different from those found in public sites using the original survey (Hirschhorn, et al., 2010). While it is difficult to compare across different surveys, reported HSR in this study also appears comparable to or greater than HSR reported by HIV patients in Western countries (Beck et al., 1999; Dang, Westbrook, Rodriguez-Barradas, & Giordano, 2012; Sullivan, Stein, Savetsky, & Samet, 2000) and South Africa (Govender, McIntyre, Grimwood, & Maartens, 2000).

The only site factor examined, patient volume, was positively correlated with overall quality ratings. This was surprising, as staffing levels were similar regardless of the number of patients seen (typically one provider and counselor at a clinic session), resulting in higher patient-to-staff ratios at larger sites and longer wait times. However, providers at higher-volume sites may be more experienced and thus more able to provide high-quality care.

Satisfaction with promptness was lower than other HSR domains, despite relatively short waiting times for most patients. This has been described in other studies of HIV care in sub-Saharan Africa (Wouters, et al., 2008), and may reflect high expectations when patients are paying for services.

Interestingly, location was not the most important factor in patients’ selection of a clinic. Respondents rated quality of service and confidentiality as most important, with convenience of location the third-most cited reason. Respondents reported work and travel schedules as the most common reasons for missing appointments. Our findings suggest that in the setting of urban private-sector clinics, flexible clinics hours, prompt services, and efforts to improve respect, privacy and confidentiality may be more helpful in increasing visit adherence than geographic accessibility, which is often cited as important in ensuring adherence (Geng, et al., 2010). Reflecting this finding, we did not find any relationship between self-reported visit adherence and travel time to the clinic or difficulty reaching the clinic (data not shown). While there are likely differences in perceptions of HSR between countries, these findings may be helpful for urban health facilities in sub-Saharan Africa interested in improving HSR.

This study has a number of limitations. PEPFAR-supported sites may not be representative of all private clinics providing HIV care in Dar es Salaam and other sites in Tanzania, limiting generalizability. It is also likely that even with a self-administered survey, some patients may be reluctant to complain. We are also unable to assess the contribution of technical quality of care to patient satisfaction. The small number of sites limited our power to identify facility-level factors other than size that may have influenced HSR. Finally, the lack of qualitative data in a brief satisfaction survey makes it difficult to offer specific recommendations to improve services.

Despite these limitations, this study demonstrates high overall HSR and areas of particular value to HIV patients receiving care at PEPFAR-supported urban private clinics in Tanzania. While other studies have demonstrated the impact of some aspects of HSR on adherence to care and treatment (Poles, et al., 2012; Schneider, et al., 2004), more work is needed to assess whether improvements in these areas can improve long-term adherence to HIV care and treatment. As decentralization of HIV care continues, further understanding how to expand public-private partnerships while ensuring quality and responsive care is an area of importance in reaching the goal of universal access.