The role of regional health systems on the waiting time inequalities in health care services: Evidences from Italy

Abstract

Inequalities in effective access to healthcare are present among countries and within the same country. Despite in Italy exist the principle of equity in access to health system, there are evidence of different access rates in the form of unequal waiting time within the country. Waiting times are an instruments to ration healthcare services dealing with resource scarsity. Theoretically, it is a fair tool because waiting times should depend only on health needs and not on the ability to pay. However, a growing literature has pointed out that belonging to a particular socioeconomic status leads to waiting times inequalities for healthcare services. Many countries have socioeconomic disparities among regions, and healthcare organizations need to take into account these differences. The increasing power of Regional Health Authorities in decentralized health systems, as in the case of Italy, has generated different organizational ways to provide health care, possibly leading to different access rates in the form of unequal waiting time within the country. This paper aims to understand if the administrative area (Regional Health Authorities) in charge of health services affects waiting times lowering or strengthening health care access inequalities. Using a series of logistic regression models, this work suggests the presence of two vectors: socioeconomic inequalities and regional inequalities. Health organizations need to implement different kinds of answers for each vectors of inequalities.

Keywords

health literacy National Health System patient empowerment waiting time

Introduction

Waiting times (WT) are one of the main issues in many countries;¹ excessive waiting for treatments may deteriorate patient’s health status and reduce treatment effectiveness² potentially, becoming a barrier in the access to health care services. It has been shown that inequalities in WT arise among European countries,³ but they are present also within the same country in the form of regional disparities. Regional differences in access to care may be particularly high in countries where competences for the organization of healthcare have been (partially) devolved to regions, as in the Italian case (Figures 1 and 2).

Figure 1.

Specialist visits—percentages of individuals reporting excessive waiting time with respect to their need by region (and 95% confidence intervals).

Figure 2.

Diagnostic tests—percentages of individuals reporting excessive waiting time with respect to their need by region (and 95% confidence intervals).

Since one of the main goals of national health systems is to provide equal access and equal treatment for equal need, patients’ waiting time has to be related only to the health need, so people with the same health need have to wait the same time, without any difference due to other factors.⁴ Nevertheless, a recent literature review showed the existence of a negative association between length of waiting time and socioeconomic status (SES).⁵ Education and income are the domains that most affect waiting times.^1,6–17 The mechanism under this phenomena needs to be explored more in detail. The literature has highlighted some possible explanations.⁵ The less affluent people may have lower ability to keep up with the system as for example not losing appointments, or obtaining more information and exercise pressure in case of undue delay.^7,11 Therefore, it seems to be much more a cultural problem, related to bad habits because people with economic deprivation and low education tend to have inadequate health literacy.^18,19 Health literacy is the individual ability “to obtain, process, and understand basic health information and services needed to make appropriate health decisions.”²⁰ It has been shown that individuals with low health literacy are significantly more likely than individuals with adequate health literacy to delay or forgo needed care or to report difficulty finding a provider.²¹ Low health literacy is more common in low income population, immigrants, people, and people with fewer years of education.²² In several countries there are areas deprived or with a concentration of people with lower income or education.

Therefore, one of the reasons for WT inequalities within a country may reside in the presence of a large part of population in low socioeconomic levels and not in a lower regional health systems effectiveness. Using the framework developed by Macintyre et al.²³ which investigates the effects of a place or an area on individuals’ health, it is possible to underline two reasons for geographic differences:

An over-concentration of individuals (classified according to social status, type of employment, minorities, etc.) with a specific socioeconomic profile, that is the compositional effect. For waiting times, it translates in people living in a certain area having an average socioeconomic status level which affects waiting because, for example, there is a higher need of health services in some areas and health services suffer for it. In fact, it has been shown that usually individuals with lower socioeconomic status may have a higher severity and morbidity compared to those with higher socioeconomic status using healthcare services more times.^28–30

Second, the presence or not of amenities and opportunity structures or the presence of efficient health services. In any specific administrative area, there are different supply and organizational characteristics leading to different wait and eventually inequalities in waiting times. This second level is the result of organizational, managerial and political decisions.

Regional differences in WT can be due to different population characteristics with related challenges due to different health needs and not to lower regional health systems effectiveness. In literature, there are two main ways of evaluating WT inequalities: one studies how different health systems impact inequalities and the second assesses the impact of SES on access to care. The aim of this study is to advance the literature exploring the possibility to merge the two points of view taking advantage of the peculiarities of the Italian case and the use of a wide national survey with individual data. The practical goal is to verify if the administrative area in charge of delivery health services may affect waiting times inequalities once controlled for two main individual factors: health status and socioeconomic status. This is an explorative study that aims to understand, keeping in mind the complexity of the phenomena, if regional health systems are able at lowering or strengthening WT socioeconomic inequalities showed in previous studies.^5–17 The objectives of this study are of particular interest for countries that rely on National Health Systems, which decentralized health responsibilities or for policy makers who need to understand where inequalities in waiting time are generated.

The novelty of this study is twofold. A wide national survey with in-depth information on patients is used. It is possible to match individual data such as health status and socioeconomic status with perceived waiting time and with the region of residence. Second, it has been possible to examine data on two different typologies of health care services specialist visits, diagnostic tests that are far less studied with regard to elective surgeries.^5,31

The results show the presence of SES inequalities in WT along with regional disparities. Even controlling for individual SES, regional disparities arises providing more evidence on organizational and managerial differences in regional health systems. Results suggest that in Italy there are two vectors of disparities, the individual SES and the regional health systems. The existence of these two dimensions implicates different answers.

Institutional setting

Italian National Health Service

The Italian National Health System (INHS) was founded in 1978 with the goal to guarantee a wide health insurance coverage and uniform health care to all citizens across the country, regardless of socio-economic characteristics and area of residence. Since its birth, the INHS has utilized, as most national systems, a decentralized model to respond better to specific need through local authorities, but in the 90s this tendency has been increasing culminating in the 2001 constitutional reform that formalized a hierarchical structure with wide power to regional authorities. INHS is organized on a three-level structure: central (Ministry of Health), regional (20 Regional Health Authorities) and local (Local Health Authorities, Aziende Sanitarie Locali ASLs). The main idea of the reform was that the regional level could have a better knowledge of local peculiarity leading to improve health care responsiveness, upgrade efficiency and contain expenditure. The regions have the political and administrative jurisdiction over the health sector (only partially the fiscal power), so they are in charge to organize their own health systems with the constraint to deliver essential levels of care (Livelli Essenziali di Assistenza, LEA) which are decided at the national level and must be guaranteed across the nation. The Italian National Health Service (Servizio Sanitario Nazionale, SSN) is grounded on the constitutional principles of universalism and equity. There is evidence that these principles are not completely reached yet.

As stated by Lega et al.,²⁴ the decentralization process of INHS has produced mixed results, as some regions took advantage of it to strengthen their systems, whereas others were not capable of developing an effective steering role. It has been noted that there are wide differences in management skills and public health capacities among regions.²⁵ Regions with weaker management skills turned out to be also financially distressed. Some of these regions needed to be included in Recovery Plans. Rosso et al.²⁶ showed associations between Recovery Plans and low quality of projects, possibly due to weak regional public health capacities leading to inequalities among regional systems. Inequalities in the ability to achieve the goal arise among regions.^25,27 The increasing power of the Regional Health Authorities has generated different organizational ways to provide health care potentially leading to different access rates in the form of unequal waiting time within the country. Although the INHS is grounded on the constitutional principles of universalism and equity, there is evidence that these principles are not completely reached, for example inequalities in WT arise among regions (Figures 1 and 2).

Data and methods

Data and study design

The empirical analysis makes use of data collected by the Italian Health Interview Survey 2013 “Multiscopo” (Health conditions and recourse to the Health Services in Italy 2013, “Condizioni di salute e ricorso ai servizi sanitari” performed by ISTAT, Italy’s National Statistics Institute) and refers to a sample of about 120,000 individuals. This work takes into account a lower number because only people who have effected at least one specialized visit (30,861) or a diagnostics test (29,606) in the latest 12 months have been selected. In this sample, we have not considered young people under 15.

The use of “multiscopo” allows us to have more information with regard to usual survey data; the sample size was wider and contains detailed information on patients’ health.

People who received the health care service in the last year had to answer this question:

Have you had to wait to do your last [visits, tests, hospitalization] with respect to the time it was necessary?

Therefore, waiting time is measured as a binary variable: “Yes, I have had to wait because I had no other choice” and “No, I have not had to wait more than necessary.” The question measures the patient self-evaluation of the acceptable time with respect to his health need, that is if he has experienced excessive waiting time with respect to his need.

There are many ways to measure WT. Usually it is the time (days, weeks) before receiving the service according to the start and end point chosen or available (i.e. out-patient waiting time = time elapsed from the date of general practitioner (GP) referral to the date of specialist assessment; in-patient waiting time = time elapsed from the specialist addition in the list to treatment or the sum of the above two that is referral-to-treatment waiting time, that is, the time elapsed between family doctor referral to treatment). Another criterion used in some studies is the notion of excessive waiting time with respect to patient’s need.^14–17

The approach starts from the fact that having a certain amount of waiting time is physiological, what is negative is excessive waiting time where patient’s health starts to be at risk of deterioration. Waiting times should be compared with a fixed time threshold based within which it is necessary to be cured to avoid a heavy deterioration in health.³⁰ The independent variables or risk predictors according to the literature review⁴ can be classified into five categories: (i) demographic, (ii) socioeconomic, (iii) health condition, (iv) structure typology, and (v) areas/regions (see Table 1).

Table 1.

Variable description, summary statistics and share of population that state that they had to wait more than necessary with respect to their need for the two services.

Variable	Specialist visits (N = 30,861)		Diagnostic tests (N = 29,606)
	N (%)	Yes, I waited, %	N (%)	Yes, I waited, %
Sex
Male	15,107 (49%)	26%	11,560 (39.0%)	25%
Female	15,574 (51%)	28%	18,046 (61.0%)	24%
Age
15–24	2206 (7.1%)	22%	1706 (5.8%)	17%
25–44	6355 (20.6%)	23%	7346 (24.8%)	21%
45–64	10,542 (34.2%)	26%	10,953 (37.0%)	25%
65–74	5738 (34.2%)	30%	5063 (17.1%)	29%
74+	6020 (18.6%)	30%	4538 (15.3%)	27%
Education
Lower than high school	14,068 (66.3%)	39.2%	17,527 (59.2%)	27.4%
High school certificate	5099 (24.0%)	33.1%	8335 (28.2%)	21.3%
University degrees	2067 (9.7%)	28.1%	3744 (12.6%)	17.9%
Professional qualification
Unemployed	14,479 (68.2%)	38.7%	15,910 (63.3%)	30.0%
Unskilled worker	2420 (11.4%)	35.5%	3156 (12.6%)	28.3%
Employee, skilled worker	3221 (15.2%)	32.6%	4618 (18.4%)	24.1%
Manager, entrepreneur	1114 (5.2%)	24.3%	1463 (5.8%)	20.2%
Macro—areas
North-West	7014 (22.7%)	23.5%	6971 (23.5%)	20.8%
North-East	7090 (23.0%)	26.1%	7096 (24.0%)	23.2%
Center	5798 (18.8%)	29.0%	5756 (19.4%)	27.8%
South	7576 (24.5%)	30.2%	6660 (19.4%)	27.7%
Islands	3387 (11.0%)	27.0%	3123 (10.5%)	22.5%
Type of structure
Public	21,234 (68.8%)	36.7%	25,147 (84.9%)	28.1%
Private	9,627 (31.2%)	6.2%	4459 (15.1%)	3.9%
Self-perceived health status
Very bad health	642 (3.0%)	36.0%	627 (2.5%)	30.8%
Bad health	2624 (12.4%)	40.3%	2646 (10.5%)	33.4%
Fair health	8099 (38.1%)	40.4%	9022 (35.9%)	31.8%
Good	8172 (38.5%)	33.4%	10,759 (42.8%)	24.9%
Very good	1697 (8.0%)	29.2%	2093 (8.3%)	21.3%
Chronic diseases
Yes	10,149 (7.8%)	39.2%	11,044 (43.9%)	31.8%
No	11,085 (2.2%)	34.3%	14,103 (56.1%)	25.3%
Limitations in life
Important limitations	3602 (11.7%)	30.7%	3044 (10.3%)	28.0%
Limitations	9059 (29.4%)	31.1%	8252 (27.9%)	29.0%
No Limitations	18,200 (59.0%)	24.5%	18,310 (61.8%)	23.0%

The education and the income levels are measured with two variables concerning respectively the school certificates or degrees and the self-evaluation of the economic resources available for the family. Other socio-economic variable is the professional qualification (occupied or not and the professional qualification). Finally, taking into account the health condition, the self-evaluation was used together with daily limitations in normal life and the presence of chronic diseases. The structure typology is formed by private and public structure. In the second group are included every structure that is funded by public sector, as NHS structure and no profit or contracted private structures. In regard of the geographic variable, we ran analyses both with single regions and macro-area and the results were not affected by this choice. To simplify the sub-national analysis, the 19 Italian regions and the 2 autonomous provinces were grouped into five macro-areas: North-West, North-east, Center, South, and Islands (that are the macro areas used by Eurostat).

Method of analysis

We entered all of the above explanatory variables into multivariate logistic regression models. We start by estimating the following model to establish the “raw” socioeconomic gradient in waiting time, that is, differences in waiting time by education level, employment conditions and economic resources conditional on age and gender.

Logit {(p_{w t})}_{i} = {\log (\frac{p_{w t}}{1 - p_{w t}})}_{i} = δ_{0} + {\vec{δ}}_{1} {DEM}_{i} + {\vec{δ}}_{2} {SES}_{i} + ε_{i}

(1)

where wt is the binary variable indicating the share of individuals that have stated to have experienced excessive waiting with respect to their health need, DEM is a vector of demographic variable, SES is a vector of socioeconomic conditions, and Ɛ is an error term.

Part of the socioeconomic gradient may reflect the fact that usually individuals with higher socioeconomic status may have a lower severity compared to those with lower socioeconomic status.¹⁶ Patients are prioritized on the waiting list; therefore, more severe patients should wait less.^6,7,10 Therefore, the lack of adequate controls on health status (in addition to age and gender) might potentially generate biased results.

We extend equation (1) in three steps to study the contributions of various pathways to the gradient in waiting time. We first control for health conditions

Logit {(p_{w t})}_{i} = {\log (\frac{p_{w t}}{1 - p_{w t}})}_{i} = δ_{0} + {\vec{δ}}_{1} {DEM}_{i} + {\vec{δ}}_{2} {SES}_{i} + {\vec{δ}}_{3} {H C}_{i} + ε_{i}

(2)

where HC is a vector of health conditions.

A comparison of equations (1) and (2) tells us how the raw socioeconomic gradient in waiting time is due to socioeconomic differences in disease patterns with the more deprived usually experiencing worse health.

Then, we investigate whether inequalities arise “across” regions or “within” regions.

We include a vector of dummy variables for areas (REG).

Logit {(p_{w t})}_{i} = {\log (\frac{p_{w t}}{1 - p_{w t}})}_{i} = δ_{0} + {\vec{δ}}_{1} {DEM}_{i} + {\vec{δ}}_{2} {SES}_{i} + {\vec{δ}}_{3} {H C}_{i} + {\vec{δ}}_{4} {REG}_{i} + ε_{i}

(3)

A comparison of equations (2) and (3) tells us how the gradient is affected by the regional system and by the individual SES to understand whether inequalities arise “across” regions or “within” the region.

Finally, we include a dummy variable indicating the use of private structures

Logit {(p_{w t})}_{i} = {\log (\frac{p_{w t}}{1 - p_{w t}})}_{i} = δ_{0} + {\vec{δ}}_{1} {DEM}_{i} + {\vec{δ}}_{2} {SES}_{i} + {\vec{δ}}_{3} {H C}_{i} + {\vec{δ}}_{4} {REG}_{i} + δ_{5} {PRIV}_{i} + ε_{i}

(4)

A comparison of equation (4) with equation (3) allows us to understand if the use or not of private sector affects the socioeconomic gradient in excessive waiting time.

We examined the issue of potential multicollinearity of education, economic resources and employments conditions calculating their variance inflation factors (VIFs). The values of VFIs are below the threshold of 4, and hence the hypothesis of multicollinearity can be rejected.

Results

Specialist visit

In 2013, 27.2% of the Italian population aged 15 and older endure excessive waiting time for specialist visits over the past 12 months. As showed in Figure 1 there are wide differences among regions and different areas of the country. The rate of self-evaluated excessive waiting varied from 20.5% in the Province of Bolzano or 20.7% in Liguria to 37.8% in Molise (Figure 1). The analysis by geographical macro-area (Table 2) provides insights that in the northern part of Italy the percentages of people stating excessive waiting time were significantly lower than the national average, with the minimum value observed in the North-West (23.5%). On the contrary, the percentage of individuals with excessive waiting time reached its maximum in the Center and southern part of the country where a value of around 30% was found. These differences can be due to a compositional effect; in fact on average, the socioeconomic level is lower in the southern regions when compared to the north of Italy. It becomes interesting to understand if controlling for socioeconomic effect inequalities arises across or only within regions.

Table 2.

Specialist visit and diagnostic tests—percentages of individuals reporting excessive waiting time by macro-area.

Macro area	Regions	Percentage	SE	CI
Specialist visits
North-West	Piemonte, Valle d’Aosta, Lombardia, Liguria	23.50%	0.005	(0.2–0.24)
North-East	Bolzano and Trento, Veneto, Friuli Venezia Giulia, Emilia Romagna	26.10%	0.005	(0.25–0.2)
Center	Toscana, Umbria, Marche, Lazio	29.00%	0.006	(0.27–0.29)
South	Abruzzo, Molise, Campania, Puglia, Basilicata, Calabria	30.20%	0.005	(0.29–0.31)
Islands	Sicilia, Sardegna	27.00%	0.008	(0.25–0.28)
Diagnostici tests
North-West	Piemonte, Valle d’Aosta, Lombardia, Liguria	20.70%	0.005	(0.20–0.22)
North-East	Bolzano and Trento, Veneto, Friuli Venezia Giulia, Emilia Romagna	24.50%	0.005	(0.23–0.26)
Center	Toscana, Umbria, Marche, Lazio	27.80%	0.006	(0.27–0.29)
South	Abruzzo, Molise, Campania, Puglia, Basilicata, Calabria	27.70%	0.006	(0.27–0.29)
Islands	Sicilia, Sardegna	22.50%	0.007	(0.21–0.24)

CI, 95% confidence intervals; SE, standard errors.

The results of multivariate logistic regression models are presented in Table 3. For reasons of clarity, the results are reported in the nomenclature of territorial units for statistics (NUTS), the official division of the EU for regional statistics (see Table 3), by aggregating regions. Using the single regions did not change the results.

Table 3.

Specialist visits—multivariate logistic regression models for self-evaluated excessive waiting time.

	OR [95% CI] (1)	OR [95% CI] (2)	OR [95% CI] (3)	OR [95% CI] (4)
Sex
Female	1.00	1	1
Male	1.06 [0.99–1.11]	1.06 [0.99–1.11]	1.05 [0.99–1.10]	1.02 [0.95–1.07]
Age
15–24
25–44	1.15 [1.01–1.30]	1.09 [0.96–1.24]	1.09 [0.97–1.25]	1,10 [0.97–1.25]
45–64	1.32 [1.17–1.49]	1.19 [1.04–1.35]	1.19 [1.05–1.35]	1.19 [1.05–1.35]
65–74	1.49 [1.31–1.69]	1.29 [1.13–1.48]	1.30 [1.14–1.48]	1.30 [1.14–1.48]
74+	1.46 [1.28–1.65]	1.25 [1.10–1.43]	1.25 [1.09–1.43]	1.25 [1.09–1.43]
Economic resources
Adequate or more
Insufficient	1.40 [1.32–1.48]	1.36 [1.29–1.44]	1.35 [1.27–1.42]	1.34 [1.26–1.41]
Completely insufficient	1.70 [1.53–1.89]	1.63 [1.47–1.82]	1.60 [1.44–1.78]	1.58 [1.42–1.76]
Education
University degrees	1
High school Diploma	1.08 [0.98–1.20]	1.07 [0.97–1.19]	1.07 [0.97–1.19]	1.07 [0.96–1.18]
Lower than high school Diploma	1.32 [1.20–1.46]	1.30 [1.17–1.43]	1.31 [1.18–1.44]	1.27 [1.15–1.41]
Health Status		X	X	X
Regions
Areas			X	X
North-West			1	1
North-East			1.17 [1.08–1.27]	1.18 [1.09–1.27]
Center			1.31 [1.21–1.42]	1.37 [1.26–1.48]
South			1.28 [1.19–1.38]	1.34 [1.24–1.44]
Islands			1.10 [1.00–1.21]	1.14 [1.04–1.26]
Structure
Public/				1
Private				0.44 [0.39–0.49]
R-quadrato di Nagelkerke	0.034	0.038	0.041	0.051

Table 4.

Diagnostic tests—multivariate logistic regression models for self-evaluated excessive waiting time.

	OR [95% CI] (1)	OR [95% CI] (2)	OR [95% CI] (3)	OR [95% CI] (4)
Sex
Female	1.00	1	1	1
Male	1.05 [0.99–1.11]	1.05 [0.99–1.11]	1.04 [0.99–1.10]	1.01 [0.95–1.07]
Age
15–24	1	1	1	1
25–44	1.42 [1.23–1.64]	1.32 [1.14–1.53]	1.33 [1.15–1.54]	1.33 [1.14–1.54]
45–64	1.69 [1.47–1.93]	1.45 [1.26–1.67]	1.46 [1.26–1.69]	1.37 [1.18–1.59]
65–74	1.83 [1.59–2.11]	1.49 [1.28–1.73]	1.50 [1.29–1.74]	1.40 [1.20–1.64]
74+	1.58 [1.37–1.83]	1.27 [1.09–1.48]	1.27 [1.09–1.48]	1.18 [1.01–1.39]
Economic resources
Adequate or more	1	1	1	1
Insufficient	1.34 [1.27–1.42]	1.30 [1.23–1.38]	1.27 [1.20–1.35]	1.23 [1.16–1.31]
Completely insufficient	1.66 [1.49–1.86]	1.53 [1.34–1.71]	1.50 [1.34–1.68]	1.44 [1.28–1.62]
Education
University degrees	1	1	1	1
High school diploma	1.13 [1.02–1.25]	1.12 [1.01–1.24]	1.12 [1.01–1.24]	1.08 [0.98–1.20]
Lower than high school diploma	1.31 [1.18–1.45]	1.28 [1.15–1.41]	1.30 [1.18–1.44]	1.20 [1.08–1.33]
Professional qualification
Unemployed	1.50 [1.30–1.73]	1.44 [1.25–1.66]	1.43 [1.24–1.64]	1.35 [1.16–1.56]
Unskilled worker	1.46 [1.25–1.70]	1.44 [1.23–1.68]	1.44 [1.24–1.68]	1.34 [1.14–1.57]
Employee, skilled worker	1.23 [1.07–1.42]	1.22 [1.06–1.41]	1.22 [1.06–1.41]	1.18 [1.02–1.37]
Manager, entrepreneur	1	1	1	1
Health status		X	X	X
Areas			X	X
North-West			1	1
North-East			1.17 [1.08–1.26]	1.17 [1.08–1.27]
Center			1.47 [1.35–1.60]	1.67 [1.53–1.82]
South			1.33 [1.23–1.45]	1.51 [1.39–1.64]
Islands			1.00 [0.91–1.11]	1.12 [1.00–1.24]
Structure				X
Public/				1.00
Private				0.10 [0.09–0.12]
R-quadrato di Nagelkerke	0.026	0.033	0.044	0.11

Model 1 shows that, conditional on age and sex, people with a higher SES have a lower odds of experience in excessive wait. Furthermore, people stating poor economic resources have a 70% increase in the odds of stating an excessive wait respect to people with adequate economic resources. Individuals with low educational level has a +32% increase in the odds of excessive wait respect to university studies.

Comparing models with and without controls for health conditions,^1,2 the effect of education and income becomes weaker, but, as the majority of the literature pointed out, the socioeconomic gradient is still present. Heterogeneity in patients’ health does not completely explain the social gradient in waiting times.

When macro-area/regions are included in the model,³ the socioeconomic gradient persists. This suggests that socioeconomic inequalities in waiting times operate within regions or macro area. At the same time, a geographical gradient exists. Differences in the odds of excessive wait were observed among geographical areas. The odds of having experienced excessive waiting time were greater for people living in the South and the Center (OR 1.37 and 1.34) than for those residing in the North-West. The differences are lower for North-East area (OR 1.18) and are not statistically significant for Islands.

Results from Model 4 show the difference of using a private structure (OR 0.44). This result is obvious, but it can be important to view how the results for other variables change. Socioeconomic gradient remains significant for all the three variables (Education, Economic resources and Professional qualification). Model 4 reports a slight increase in macro-areas odds, in particular for Center and South macro-areas, from 1.31 to 1.37 and from 1.28 to 1.34, respectively. The changes are wider for diagnostic tests (see Table 3).

Diagnostic tests

In 2013, 24.5% of the Italian population aged 15 and older experienced excessive waiting time for health diagnostics tests over the past 12 months (Figure 2). The rate of self-evaluated excessive waiting varied from 16.4% in the Province of Bolzano to 34.1% in Molise.

The analysis by geographical macro-area (Table 2) provides insights that in the northern part of Italy the percentages of people stating excessive waiting time were significantly lower than the national average, with the minimum value observed in the North-West (20.7%). On the contrary, the percentage of individuals with excessive waiting time reached its maximum in the Center and southern part of the country where a value of 27% was found (see Figure 2).

The results of multivariate logistic regression models are presented in Table 4.

Model 1 shows that, conditional on age and sex, people with a higher SES have a lower odds of experience excessive wait. People stating poor economic resources have 66% increase risk ok excessive wait respect to people with adequate economic resources, while who has a low educational level has an increase of 31% respect to university studies.

Comparing models with and without controls for health conditions,^1,2 the effect of education and income becomes lightly weaker, but the socioeconomic gradient is still present. Heterogeneity in patients’ health does not completely explain the social gradient in waiting times.

When macro-area/regions are included in the model,³ the socioeconomic gradient persists. Here SES coefficients can be interpreted as differences in waiting time by SES level within regions for patients with the same demographic characteristics and health conditions. Therefore, results suggest that socioeconomic inequalities in waiting times operate within regions or macro area. At the same time, a geographical gradient seems to exist. Differences in the odds of excessive wait were observed among geographical areas. The odds of having experienced excessive waiting time were greater for people living in the South and the Center (OR 1.33 and 1.47) than for those residing in the North-West. The differences are lower for North-East area (OR 1.17) and are not statistically significant for Islands.

Results from Model 4 show the importance of using a private structure. Socioeconomic gradients become flatter for all the three variables (Education, Economic resources, and professional qualification), but they remain significant. High school diploma and university studies are not statistically significant and lower education level have reduced odds (OR 1.30 vs. 1.20). Complete and insufficient economic resources have lower odds, but it remains significant (OR 1.44 and 1.23) as professional qualification. Interestingly, Model 4 reports changes in macro-areas odds, in particular an increase for Center, South, and Islands macro-areas. For example, the Center and South areas odds increase from 1.47 to 1.67 and from 1.33 to 1.51. This could mean that the composition of the supply side is different in the areas and has its own effect on waiting time.

Discussion

INHS, as in other countries, has decentralized the organizations and policy power to regional systems. Even if Regional Health Authorities must reach common goals defined at the national level (LEA), the decentralization process of INHS has produced mixed results, as some regions took advantage of it to strengthen their systems, whereas others were not capable of developing an effective steering role²⁴–25 leading to differences in performance as for example access to care.²⁷ As shown in the descriptive statistics there are wide differences among regions and areas of the country in waiting time for specialistic visits and diagnostics. An increasing literature is showing how waiting time is affected by SES of the individuals or the SES of the area of residence.^5–17 Since in Italy, as in other countries, a socioeconomic gradient exists among regions, this work analyzed if these differences can be due to compositional effect or arise also due to the related contextual factors. The regression models allowed us to understand if controlling for socioeconomic effect inequalities arises across or only within regions.

First, we show that for the health services analyzed socioeconomics inequalities regard both education and income. Second, the gradient, as expected is affected by the inclusion of detailed controls for severity and health need as captured by several health proxies (chronicity, self-reported health status, limitations in daily activities). Often socioeconomic disparities are linked at inequalities in health status with the most deprived having worse health and sometimes entering the health systems later with major need.^28–30 Under such a hypothesis, we would expect a substantial fall in the overall SES gradient after controlling for health status, which instead we do not observe. This confirms evidence of previous studies that reported a significant socioeconomic gradient in waiting times even after controlling for health status.^5–17

Inequalities arise both “within” and “across” regions and macro-areas. Residents in Center or South Italy have a higher likelihood to experience excessive waiting time with respect to Italians living in the North. Results suggest that belonging to different regional systems leads to different probabilities of waiting more than necessary. Still individuals’ socioeconomic disparities within same areas play a role in affecting waiting time. The analyses showed that waiting time inequalities are affected by individual socioeconomic status but also by the regional health system of belonging.

The last point addressed was the type of structure used. As expected, people who used private structures wait far less than who used public sector. This is true in particular for diagnostics tests. Interestingly, controlling for the type of structure, socioeconomic gradient becomes flatter but still is significant and the macro-areas disparities extend. Including the type of structure in the model increases the inequalities gradient across regions and lower it within. This confirms the hypothesis that wealthier people use the public sector if the time waited is acceptable, otherwise they are able to jump the queue using the private structures upon payment. The widening areas disparities can be due to a major presence of the private sector in the Center and the South of Italy.

One of the strengths of this study lies in the use of a unique survey dataset. Usually these types of studies use administrative or survey data. Survey data have the advantage of matching accurately the patient’s waiting times with his income, his education and health conditions. Usually, in literature, socioeconomic status is assigned to the individual record according to his postal code or address linking this information with small-area deprivation indexes calculated with census data. The dataset used in this study collects very extensive patients’ severity variables, typically available only for administrative data and at the same time several individual socioeconomic status variables, usually available only in survey data. This allowed enhancing the knowledge on the association between waiting time and socioeconomic status controlling accurately for patients’ severity. Moreover, it has been possible to examine data on two different typologies of health care services (specialist visits and diagnostic tests) that are far less studied respect to elective surgeries.^5,32

Conclusions

This empirical work suggests that, in spite of the INHS statutory obligation to provide equal access according to the needs of all Italian citizens, socioeconomic and geographical inequalities in waiting time exist. The innovation of the work is the attempt to highlight how the way regional health systems are organized and managed can have a concrete impact on inequalities on healthcare access. For example, individual inability to keep up with the system may be caused by individual low health literacy and healthcare organizations need to play a role becoming aware of SES inequalities in health care access finding ways to minimize these inequalities. In line with previous international studies^5–17 logistic regression results suggest that more vulnerable population groups experience longer waiting times for primary health care. The regional WT differences were expected to reduce or disappear once controlled for SES, but this is not the case. Therefore, regional differences are not due, or not completely due, only to a compositional effect, that isthe concentration of population with low income and low education. Results suggest that in Italy there are two vectors of disparities, the individual SES and the regional health system. The existence of these two dimensions implicates different answers.

First, from an economic and policy viewpoint, waiting times inequalities depend not only on resources devoted to healthcare or to managerial solutions but also to those assigned to social determinants and education. Improving general socioeconomic conditions, in particular attacking cultural determinants and education,²⁹ could affect positively waiting times inequalities and the ability to keep up with the system may improve. From the point of view of health policy, future research may explore the supply side characteristics for each region, as for the number of public/private structures, the number of physicians, economic efficiency of the structures. On the other hand, we need to take into account the wide differences in management skills and public health capacities among regions.²⁵

Second, it is important to implement practical and managerial interventions able to have effects in a shorter time. Health systems should be able to weaken access inequalities and not strengthen them. Starting from the hypothesis presented in the introduction, the interaction between patient and the system seems to be a key point. In healthcare services, the relationship between clinicians and patients is crucial. The clinician has knowledge of diagnosis, treatment options, and prognosis, and the client knows about the experience of illness, social circumstances, and personal preferences.³² The contribution of each participant is important. On the other hand, limited health literacy, usually showed by people with low SES,^18,19 inhibits patient empowerment.³² Limited health literacy is the gap between the patients and the healthcare professionals, which prevents them to establish a co-creating relationship. The patients’ health literacy is not the only variable to affect the relationship; it strongly depends on the complexity of the health care service system, as well as on the communication skills of health care providers that is the organization health literacy.¹⁹ Most of the literature defines organizational health literacy as the capacity of the health care organizations to establish clear and comfortable relationships with the patients and the ability of health care providers to address the information needs of people living with limited health literacy skills.³³ Organizational changes need to enhance the interactions between the providers of care and the patients. Health care organizations have to: (i) make health literacy integral part of their mission, structure, and operations; (ii) implement strategic educational initiatives to prepare their health care professionals, in order to make them aware about the impacts of poor health literacy on the outcomes of health care provision; (iii) include it into strategic and operational planning, evaluation activities, patient safety concerns, and quality measures; (iv) engage the served population in the design, implementation, and evaluation of health information and services to identify the knowledge needs of people; and (v) enhance the comprehensibility of health information they provide, designing and distributing print, audiovisual, and social media.³³

This paper has some limitations. First, the waiting time measure is a subjective evaluation. In our study, we use the concept of excessive WT self-reported by the patient. We use the term excessive because the real issue is not (only) the physiological presence of waiting times but the excessive waiting time where patient’s health starts to be at risk of deterioration. Waiting times should be compared with a fixed time threshold-based clinical evidence within which it is necessary to be cured to avoid a heavy deterioration in health. For our study it would have been better having a threshold based on clinical evidence. On the other hand, Gaudet in a work on total hip replacement showed that the patients’ valuations of a threshold for reasonable waiting were similar to the one stated by specialists.³⁴ Therefore, we use the variable as a proxy of excessive WT. A second limitation is that the National survey design is cross-sectional, and, thus, data on outcomes and determinants are collected simultaneously; therefore, we cannot draw definitive casual relations. Further research may be devoted, firstly, to unpack what causes inequalities in waiting times, in particular to investigate how such inequalities are due to patients’ ability to interact with the systems and to understand practical ways to reduce inequalities in health care service access.

Second, the analysis should be extended to an ad hoc quantitative survey administered on patients to reveal their health literacy levels and their attitude and behavior with regard to WT.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Stefano Landi

References

Siciliani

Moran

Borowitz

Measuring and comparing health care waiting times in OECD countries. Health Policy. 2014; 118: 292–303.

Koopmanschap

Brouwer

WBF

Hakkaart-van Roijen

, et al. Influence of waiting time on cost-effectiveness. Soc Sci Med. 2005; 60: 2501–2504.

Baeten

Spasova

Vanhercke

, et al. Inequalities in access to healthcare. A study of national policies. Brussels: European Social Policy Network (ESPN), European Commission, 2018. http://www.ose.be/FR/ publications.htm.

Valente

Testi

Tanfani

, et al. A model to prioritize access to elective surgery on the base of clinical urgency and waiting time, BMC. Health Serv Re. 2009; 9: 334–351.

Landi

Ivaldi

Testi

Socioeconomic status and waiting times for health services: An international literature review and evidence from the Italian National Health System. Health Policy. 2018; 122; 334–351. DOI: 10.1016/j.healthpol.2018.01.003.

Schulz

The intertwined relationship between patient education, hospital waiting times and hospital utilization. Health Serv Manage Res. 2017; 30: 213–218.

Siciliani

Verzulli

Waiting times and socioeconomic status among elderly Europeans: evidence from share. Health Econ. 2009; 18: 1295–1306.

Saunders

Lee

Alexander

, et al. Racial disparities in reaching the renal transplant waitlist: is geography as important as race? Clin Transplant. 2015; 29: 531–538. Jun

Smirthwaite

Lundström

Wijma

, et al. Inequity in waiting for cataract surgery–an analysis of data from the Swedish National Cataract Register. Int J Equity Health. 2016; 15: 10.

10.

Petrelli

De Luca

Landriscina

, et al. Socioeconomic differences in waiting times for elective surgery: a population-based retrospective study. BMC Health Serv Res. 2012; 12: 268.

11.

Laudicella

Siciliani

Cookson

Waiting times and socioeconomic status: evidence from England. Soc Sci Med. 2012; 74: 1331–1341.

12.

Carlsen

Kaarboe

OM.

The relationship between educational attainment and waiting time among the elderly in Norway. Health Policy. 2015; 119: 1450–1458.

13.

Bouche

Ingrand

Mathoulin-Pelissier

, et al. Determinants of variability in waiting times for radiotherapy in the treatment of breast cancer. Radiother Oncol. 2010; 97: 541–547. Dec

14.

Abasolo

Negrin-Hernandez

Pinilla

Equity in specialist waiting times by socioeconomic groups: evidence from Spain. Eur J Health Econ. 2014; 15: 323–334.

15.

Barone

Fusco

Colais

, et al. Wait for surgery after hip fracture. Int J Qual Health Care. 2009; 21: 379–386.

16.

Szynkaruk

Stephens

Borschel

, et al. Socioeconomic status and wait times for pediatric surgery in Canada. Pediatrics. 2014; 134: e504–511.

17.

Hacker

Stanistreet

Equity in waiting times for two surgical specialties: a case study at a hospital in the North West of England. J Public Health (Oxf). 2004; 26: 56–60.

18.

Palumbo

Annarumma

Adinolfi

, et al. The Italian Health Literacy Project: insights from the assessment of health literacy skills in Italy. Health Policy. 2016; 120: 1087–1094.

19.

Palumbo

Designing health-literate health care organization: a literature review.

Health Serv Manage Res. 2016; 29: 79–78.

20.

Baker

DW.

The meaning and the measure of health literacy. J Gen Intern Med. 2006; 21: 878–883.

21.

Levy

Janke

Health Literacy and Access to Care. J Health Commun. 2016; 21: 43–50. DOI: 10.1080/10810730.2015.1131776.

22.

Greenhalgh

Health literacy: towards system level solutions. BMJ. 2015; 350: h1026. DOI: 10.1136/bmj.h1026.

23.

Macintyre

Ellaway

Cummins

Place effects on health: how can we conceptualise, operationalise and measure them?

Soc Sci Med. 2002; 55: 125–139. DOI: 10.1016/S0277-9536(01)00214-3).

24.

Lega

Sargiacomo

Ianni

The rise of governmentality in the Italian National Health System: physiology or pathology of a decentralized and (ongoing) federalist system?

Health Serv Manage Res. 2010; 23: 172–180.

25.

Ferrè

Cuccurullo

Lega

The challenge and the future of health care turnaround plans: evidence from the Italian experience. Health Policy. 2012; 106: 3–9. DOI: 10.1016/j.healthpol.2012.03.007.

26.

Rosso

De Vito

Marzuillo

, et al. The negative effect of financial constraints on planning prevention activities: some evidence from the Italian experience. Eur J Public Health. 2015; 25: 1117–1119. DOI: 10.1093/eurpub/ckv087.

27.

Cavalieri

Geographical variation of unmet medical needs in Italy: a multivariate logistic regression analysis. Int J Health Geogr. 2013; 12: 27.

28.

Bruzzi

Ivaldi

Landi

Non-compensatory aggregation method to measure social and material deprivation in an urban area: relationship with premature mortality. Eur J Health Econ. 2020 Apr; 21(3): 381–396. DOI: 10.1007/s10198-019-01139-x.

29.

Mackenbach Johan

The persistence of health inequalities in modern welfare states: the explanation of a paradox. Soc Sci Med. 2012; 75: 761–769.

30.

Lostao

Regidor

Geyer

, et al. Patient cost-sharing and social inequalities in access to health care in three Western European countries. Soc Sci Med. 2007; 65: 367–376.

31.

Landi

Ivaldi

Teti

Socioeconomic status and waiting times for health services: current evidences and next area of research. Health Serv Insights. 2019; 12: 1–3. DOI: 10.1177%2F1178632919871295.

32.

Palumbo

Examining the impacts of health literacy on healthcare costs. An evidence synthesis. Health Serv Manage Res. 2017; 30: 197–212. DOI: 10.1177/0951484817733366.

33.

Brach

Keller

Hernandez

, et al. Attributes of a health literate organization. Washington, DC: National Academies Press, 2012.

34.

Gaudet

Feldman

Rossignol

, et al. The wait for total hip replacement in patients with osteoarthritis. Can J Surg. 2007; 50: 101–109.