Table 4. Results of exploratory factor analysis and of overall satisfaction models, Odisha, India, 2020.

Category and experience item	Mean item value (SE)	Exploratory factor analysis and item-level testing			Coefficient, by level
					Model Ib			Model IIc			Model IIId
		Item uniqueness	Cronbach’s αa		Item	Category		Item	Category		Item	Category
Interpersonal care from nurses (λ: 3.5)e
Courtesy and respect	3.4 (0.034)	0.221	0.785		0.65***	0.76***		0.58***	0.69***		0.59***	0.70***
Listen carefully	3.4 (0.032)	0.218	0.781		0.79***			0.74***			0.75***
Explain	3.3 (0.036)	0.371	0.780		0.81***			0.75***			0.77***
Interpersonal care from doctors (λ: 1.9)e
Courtesy and respect	3.5 (0.031)	0.359	0.785		0.91***	0.82***		0.84***	0.74***		0.86***	0.76***
Listen carefully	3.3 (0.033)	0.556	0.779		0.82***			0.73***			0.75***
Explain	3.3 (0.033)	0.319	0.785		0.72***			0.65***			0.66***
Hospital environment (λ: 1.7)e
Room clean	2.9 (0.040)	0.293	0.798		0.47***	0.33***		0.39***	0.25**		0.38***	0.23**
Quiet	2.5 (0.044)	0.287	0.807		0.18***			0.10			0.08
General experience (λ: 1.3)e
Talk about pain	2.6 (0.056)	0.445	0.790		0.90***	0.68***		0.81***	0.60***		0.87***	0.62***
Talk about pain treatment	2.9 (0.036)	0.310	0.786		0.64***			0.56***			0.57***
Explain medication purpose	2.8 (0.055)	0.330	0.802		0.50***			0.42***			0.42***
After discharge (λ: 1.3)e
Assessment of post-discharge	0.9 (0.022)	0.345	0.811		0.26*	0.54**		0.09	0.34**		0.09	0.33**
Receipt discharge guidance	1.6 (0.017)	0.542	0.801		0.81***			0.59***			0.57***
Understanding of care (λ: 1.1)e
Taking preferences seriously	3.6 (0.024)	0.300	0.804		0.69***	0.69***		0.55***	0.58***		0.54***	0.57***
Understand responsibilities	3.6 (0.023)	0.171	0.801		0.72***			0.59***			0.58***
Understand purpose of medications	3.6 (0.026)	0.277	0.804		0.67***			0.59***			0.58***

SE: standard error; *P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001.

^a A typical exclusion threshold for α coefficient is 0.70. The higher the α coefficient, the more the items have shared covariance and may measure the same underlying concept. Highly correlated items will also produce a high coefficient and can therefore be interpreted as a sign of redundancy. As we did not conduct the analysis to shorten the Hospital Consumer Assessment of Health Providers and Systems survey, we retain all items regardless of performance.

^b Model I represents the unadjusted results of a bivariate ordinary least square regression where overall satisfaction is the dependent variable and each row represents a different patient experience item posed to patient.

^c Adjusted for patient age, gender and clinical complexity.

^d Adjusted for Model II factors plus interview characteristics.

^e Eigenvalues (λ) shown for retained factors. Corresponding item categories are discrete and align with factor loadings most relevant to defining each factor’s dimensionality.

Note: we excluded two items (bathroom help and explanation of medicine side-effects) from this table because fewer than 50 respondents needed support with the bathroom or were prescribed medicines.