Descriptive statistics
Tables
1 and
2 present the socio-economic and demographic characteristics of respondents.
Table 1
Socio-economic and demographic characteristics of all patients
Sex | 817 | | | |
Male | | 334 | 40.6 | |
Female | | 484 | 59.4 | |
Marital status | 810 | | | |
Unmarried | | 258 | 31.9 | |
Married | | 552 | 68.1 | |
Educational | 817 | | | |
None | | 193 | 23.6 | |
Primary-JHS | | 351 | 43 | |
Sec/Tech/Vocational | | 153 | 18.7 | |
Tertiary | | 120 | 14.7 | |
Income | 805 | | | |
GH¢100 and below | | 211 | 26.2 | |
GH¢ 101 and above | | 337 | 41.9 | |
No earnings | | 257 | 31.9 | |
Age | 793 | | | 35.5 |
Household size | 723 | | | 5.40 |
Table 2
Socio-economic and demographic characteristics of insured and uninsured patients
Sex |
Male | 195 | 36 | | 139 | 51 | |
Female | 349 | 64 | | 135 | 49 | |
Marital status |
Unmarried | 151 | 28 | | 107 | 39 | |
Married | 389 | 72 | | 163 | 61 | |
Educational |
None | 121 | 22 | | 72 | 26 | |
Primary-JHS | 242 | 45 | | 109 | 40 | |
Sec/Tech/Vocational | 100 | 18 | | 53 | 19 | |
Tertiary | 80 | 15 | | 40 | 15 | |
Income |
GH¢100 and below | 226 | 42 | | 98 | 36 | |
GH¢ 101 and above | 154 | 28 | | 82 | 30 | |
No earnings | 163 | 30 | | 94 | 34 | |
Age | | | 35 | | | 37 |
Household size | | | 5.25 | | | 5.7 |
From Table
1, out of the 817 respondents, a greater proportion (68 %) were females. With respect to marital status, a higher proportion of the respondents were married, while the least proportion (31.9 %) were unmarried. The results further showed that close to one-fourth: 193(24 %) had no formal education, 120 (15 %) had tertiary level education. A higher proportion: 239 (29 %) had middle/JHS education. The mean age of respondents was 36 years respectively. The mean household size of respondents was 5.4.
Table
2 shows that whereas the proportion of insured females were higher (64 %) compared with insured males (36 %), the proportions were almost the same in the case of uninsured females and males (49 % and 51 % respectively). The results also indicate that among those married, the insured were more (72 %) compared with the uninsured (61 %), whilst those who were not married 28 and 39 % respectively. With respect to educational background, there were no significant differences between the insured and uninsured regarding the proportions of respondents who completed the various levels of education. Twenty-two percentage (22 %) of the insured and 26 % of the uninsured had no formal education; 45 % of the insured and 40 % of the uninsured had primary/junior high school level education; 18 % of the insured and 19 % of the uninsured had secondary/vocational/technical level education; and 15 % respectively of the insured and uninsured had tertiary level education. Among those who earned monthly incomes of GHC100 ($50) or below, the insured earned a relatively higher proportion (42 %) compared with the uninsured (36 %). However, among those who earned GHC101 ($51) or above, both insured and uninsured earned similar proportions of 28 and 30 % respectively. The proportions were also similar with those who received no monthly earnings (30 and 34 % respectively). There were also no marked variations in the ages and household sizes of insured and uninsured respondents. The mean age of the insured was 35, whereas that of the uninsured was 37. On the other hand, the mean household size of the insured was 5.25; whereas those of the uninsured was 5.7.
Factor analysis of the quality of care
Table
3 presents results of the factor analysis.
Table 3
Factor analysis of the quality of care scale
Financial access |
NHIS pays the cost of all treatment | .721 | | | | | |
Cost of services are affordable | .711 | | | | | |
Exempted patients are treated free of charge | .708 | | | | | |
Only official fees are charged | .570 | | | | | |
Fairness of care |
Staff treat all patients fairly | | .764 | | | | |
Quality of drugs is same for all patients | | .763 | | | | |
Patients are treated on first-come-first-served basis | | .620 | | | | |
Very ill patients are treated first | | .586 | | | | |
Adequacy of resources & services |
Doctors are sufficient | | | .784 | | | |
Supplies are sufficient | | | .739 | | | |
Rooms in OPD are sufficient | | | .673 | | | |
Waiting time is reasonable | | | .651 | | | |
Drugs are available | | | .473 | | | |
Effectiveness of treatment |
Pharmacy instructions are clear | | | | .585 | | |
Treatment is effective for recovery and cure | | | | .501 | | |
Quality drugs are given to patients | | | | .445 | | |
Technical care |
Patients are told diagnosis | | | | | .738 | |
Patients are physically examined | | | | | .615 | |
Lab. and other tests are done | | | | | .610 | |
Patients are involved in their care | | | | | .557 | |
Interpersonal care |
Staff show compassion & support to patients | | | | | | .865 |
Staff are polite & respectful to patients | | | | | | .861 |
The initial scale used in data collection had 28 items divided into five main dimensions. However, the factor analysis reduced the items to 22, divided into six dimensions, namely, financial access to care, fairness aspects of care, adequacy of resources, effectiveness of treatment for recovery and cure, technical aspects of care, and interpersonal aspects of care.
The scale was tested for reliability. It had an overall Cronbach’s alpha value of 0.79, while the subscales ranged from 0.58 to 0.84. Thus, the reliability was highest for ‘interpersonal aspects of care’ (0.84) and lowest for ‘technical aspects of care’ (0.58). The overall mean score was 89.11, while the standard deviation was 11.457. Respondents could express their perceived quality of care on a five-point Likert scale: strongly disagree (1), disagree (2), neutral (3), agree (4), strongly agree (5).
Assessment of levels of quality of care
For purposes of assessing the levels of perceived quality of care, the mean ratings of the various indicators under a dimension was added to obtain a total quality care index for that dimension, for both the insured and uninsured patients. In the case of financial access, for instance, there are four indicators of quality of care. Based on the 5-point Likert scale, if all ratings of the indicators for this dimension were to be 1 (strongly disagree), then the total rating for financial access would be 4 (i.e. 1 × 4). If all ratings were 5, then the total rating would be 20 (i.e., 5 × 4). For ease of interpretation, the study also considered mean aggregate ratings between 4 and 12 as unfavourable or low quality; ratings between 12.01 and 16 as fairly favourable or average quality ratings; and ratings between 16.01 and 20 as favourable or good quality rating. The same approach was followed in determining the overall perception of quality of care, where ratings between 24 and 72 were considered unfavourable, 72.01 to 96, fairly favourable and 96.01 to 120, favourable.
Perceptions of quality of care between insured and uninsured
An independent sample
t-test was used to compare differences in perceptions of quality of care between the insured and uninsured. Table
4 shows the results of the comparison.
Table 4
T-test on perceived differences in quality of care between insured and uninsured patients
Financial access |
NHIS pays the cost of all treatment | 3.85 | 1.16 | 3.48 | 1.16 | 4.073*** |
Cost of services are affordable | 3.52 | 1.11 | 3.20 | 1.24 | 3.522*** |
Exempted patients are treated free of charge | 3.97 | 1.06 | 3.52 | 1.09 | 5.396*** |
Only official fees are charged | 3.70 | 1.15 | 3.89 | 1.04 | −2.376** |
Fairness of care |
Staff treat all patients fairly | 3.40 | 1.36 | 3.49 | 1.32 | 0.863 |
Quality of drugs is same for all patients | 3.37 | 1.30 | 3.40 | 1.30 | 0.316 |
Patients treated on first-come-first-served basis | 3.92 | 1.34 | 3.97 | 1.28 | 0.487 |
Very ill patients are treated first | 4.26 | 1.13 | 4.22 | 1.12 | −0.457 |
Adequacy of resources & services |
Doctors are sufficient | 2.79 | 1.30 | 2.92 | 1.28 | −1.303 |
Supplies are sufficient | 3.14 | 1.13 | 3.04 | 1.17 | −1.136 |
Rooms in OPD are sufficient | 3.24 | 1.31 | 3.24 | 1.29 | 0.080 |
Waiting time is reasonable | 2.75 | 1.37 | 2.94 | 1.30 | −1.897* |
Drugs are available | 3.29 | 1.27 | 3.33 | 1.23 | 0.354 |
Effectiveness of treatment |
Pharmacy instructions are clear | 4.56 | 0.63 | 4.51 | 0.65 | 0.988 |
Treatment is effective for recovery and cure | 4.14 | 0.86 | 3.99 | 0.85 | 2.349** |
Quality drugs are given to patients | 3.97 | 0.91 | 3.92 | 0.86 | 0.714 |
Technical care |
Patients are told diagnosis | 3.38 | 1.48 | 3.30 | 1.44 | 0.691 |
Patients are physically examined | 3.92 | 1.24 | 3.75 | 1.35 | 1.745* |
Lab. and other tests are done | 3.92 | 1.26 | 3.64 | 1.35 | 2.838** |
Patients are involved in their care | 3.68 | 1.32 | 3.71 | 1.22 | −0.274 |
Interpersonal care |
Staff show compassion & support to patients | 4.08 | 0.97 | 4.19 | 0.89 | −1.519 |
Staff are polite & respectful to patients | 4.12 | 0.98 | 4.19 | 0.91 | −1.008 |
Overall perceived quality of care | 89.48 | 11.263 | 88.38 | 11.836 | −1.212 |
The study reveals that overall, there is no significant difference in perception of quality of care between insured and uninsured patients (Insured: M = 89.48, SD = 11.263; Uninsured: M = 88.38, SD = 11.84) p = 0.226. Regarding the level of quality of care, both insured and uninsured respondents gave a fairly favourable rating, indicating that overall, quality of care in Ghana’s hospitals is somewhat good.
With respect to the individual indicators of quality of care however, there is a significant difference in perceptions of quality of care between the insured and uninsured in respect of all indicators of financial access to care (Insured: M = 3.85, SD = 1.16; Uninsured: M = 3.48, SD = 1.16) p < .001; costs of services being affordable (Insured: M = 3.52, SD = 1.11; Uninsured: M = 3.20, SD = 1.24) p < .001; and exempted patients treated free of charge (Insured: M = 3.97, SD = 1.06; Uninsured: M = 3.52, SD = 1.09) p < .001; and payment of unofficial fees (Insured: M = 3.70, SD = 1.15; Uninsured: M = 3.89, SD = 1.04) p < .018. In view of the special significance attached to financial access to health care, a multiple regression analysis was further performed to determine the association of respondents’ health insurance status on financial access to health care. The dependent variable was financial access, with insurance as the main independent variable, controlling for socio-demographic and hospital characteristics. The results indicate that insurance status is a significant predictor of financial access to health care (B = .805, p = .002). This suggests that for every additional person enrolled with the Ghana health insurance scheme, access to health care is increased by .805 points, holding other variables fixed. With regard to the levels of quality, both insured and uninsured respondents rated financial access to care fairly favourable (somewhat desirable) for all indicators.
Concerning respondent’s perceptions of fairness of care, there is no significant difference between the insured and uninsured in all indicators of quality. Three out of the four indicators of fairness of care were rated fairly favourably, while the indicator ‘Very ill treated first’, was rated favourably.
With respect to adequacy of resources, there is no significant difference in perceptions of quality between insured patients in all indicators except waiting time at 10 % significance level (Insured: M = 2.75, SD = 1.37; Uninsured: M = 2.94, SD = 1.30) p < 058. This implies that insured patients perceive waiting time to be longer, compared with uninsured patients. Two indicators, that is, adequacy of doctors and waiting time were rated unfavourably, and the remaining three indicators were rated fairly favourably.
On effectiveness of treatment, only the indicator ‘treatment is effective for recovery and cure’ shows a significant difference in perceptions of quality between insured and uninsured patients (Insured: M = 4.14, SD = 0.86; Uninsured: M = 3.99, SD = 0.85) p < 019. This suggests that more insured respondents considered the treatment they received from the hospital to be effective for recovery and cure. Regarding the levels of quality, two of the three indicators of effectiveness of treatment were rated favourably by respondents, while the indicator ‘Quality drugs are given to patients’ was rated fairly favourably by both insured and uninsured patients.
Regarding technical aspects of care, the indicator ‘laboratory and other tests are done’ shows a significant difference in perceptions of quality between insured and uninsured patients. The implication is that compared with uninsured patients, the treatment of insured patients is based more on confirmation of laboratory and other tests. At 10 % significant level, there was also a significant difference between insured and uninsured patients in respect of ‘Patients are physically examined’ (Insured: M = 3.92, SD = 1.24; Uninsured: M = 3.75, SD = 1.35) p < 082. On levels of quality of care, however, both insured and uninsured patients gave fairly favourable ratings of the technical aspects of care.
Finally, there is no significant difference between insured and uninsured patients regarding interpersonal aspects of care. However, both insured and uninsured respondents rated interpersonal aspects of care favourably.
Association between health insurance and quality of care
Beyond evaluating differences in perceptions of quality of care between insured and uninsured patients, the study further used a multiple regression analysis to examine the association between health insurance and overall perception of quality of care, controlling for socio-demographic and hospital characteristics. Table
5 presents the results of the regression analysis.
Table 5
OLS multiple regression of the association of health insurance status on perceived quality of care
(Constant) | 64.115 | |
Insurance status (Insured = 1) | 1.065 (.927) | .044 |
Age of respondent | .067* (.031) | .083 |
Marital status (Married = 1) | .244 (.949) | .010 |
Sex (Male = 1) | −.922 (.923) | −.040 |
Educational level (None = 1) | −.306 (.292) | −.043 |
Income level (No earnings = 1) | −.330 (170) | −.076 |
Distance to hospital in Km. | .007 (.056) | .005 |
Health status | 1.542** (.466) | −.215 |
Number of doctors | −1.699** (.493) | −.139 |
Size of household | .309 (.167) | .074 |
Bed capacity | .036*** (.008) | .287 |
Ownership of hospital (Government hospitals = 1) | 3.401 (.595) | .247 |
Region (Upper East = 1) | 4.102*** (.600) | .295 |
From Table
5, the results show that there is no significant relationship between insurance status and respondents’ perceptions of quality of care (B = 1.065,
p = .251). In other words, the insurance status of patients in Ghanaian hospitals is not a predictor of perception of quality of care. For the control variables, however, there was a significant relationship between age of respondents and perceptions about the quality of care (B = .067,
p = .033). The results show that an increase in respondents’ age by one year will result in improved perceptions of quality of care by .067 points, holding other variables fixed. This implies that as patients mature in age, they are more positive in their perceptions of quality of care. In order to confirm the linearity between the age of respondents and the dependent variable, age was squared in the regression model and analysed further. The results indicate that both age and age squared are not significant predictors of perceived quality of care. This suggests that age has a linear relationship with perceived quality of care. Other variables that have a positive and significant relationship with perceptions of quality of care include health status (
p = .001), number of doctors (
p = .001), hospital size (
p = .001), ownership of hospitals (
p < .001), and Region (
p < .001).