If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Department of International Health, Johns Hopkins Bloomberg School of Public Health, Baltimore, MarylandDepartment of Surgery, Columbia University, New York, New YorkSurgeons OverSeas, New York, New York
Department of Surgery, University of Washington, Seattle, WashingtonDepartment of Surgery, Komfo Anokye Teaching Hospital, Kumasi, GhanaSchool of Medical Sciences, Kwame Nkrumah University of Science and Technology, Kumasi, GhanaDepartment of Interdisciplinary Health Sciences, Stellenbosch University, Cape Town, South Africa
Although two billion people now have access to clean water, many hospitals in low- and middle-income countries (LMICs) do not. Lack of water availability at hospitals hinders safe surgical care. We aimed to review the surgical capacity literature and document the availability of water at health facilities and develop a predictive model of water availability at health facilities globally to inform targeted capacity improvements.
Methods
Using Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines, a systematic search for surgical capacity assessments in LMICs in MEDLINE, PubMed, and World Health Organization Global Health Library was performed. Data regarding water availability were extracted. Data from these assessments and national indicator data from the World Bank (e.g., gross domestic product, total health expenditure, and percent of population with improved access to water) were used to create a predictive model for water availability in LMICs globally.
Results
Of the 72 records identified, 19 reported water availability representing 430 hospitals. A total of 66% of hospitals assessed had water availability (283 of 430 hospitals). Using these data, estimated percent of water availability in LMICs more broadly ranged from under 20% (Liberia) to over 90% (Bangladesh, Ghana).
Conclusions
Less than two-thirds of hospitals providing surgical care in 19 LMICs had a reliable water source. Governments and nongovernmental organizations should increase efforts to improve water infrastructure at hospitals, which might aid in the provision of safe essential surgical care. Future research is needed to measure the effect of water availability on surgical care and patient outcomes.
Since 1990, nearly two billion people have gained access to improved sources of water, one of the success of the United Nations Millennium Development Goals.
Lack of access to water, sanitation, and hygiene (WASH) infrastructure has a significant and negative impact on both health and healthcare provision, including surgical care.
Emergency and essential surgical care: Global Initiative for Emergency and Essential Surgical Care (GIEESC). WHO Web site. www.who.int/surgery. Accessed January 30, 2016.
Much of the increased access to water has been achieved by providing piped water on-site (i.e., running water), building public taps, drilling boreholes, and supporting rainwater capture methods.
In low- and middle-income countries (LMICs) with a funded institutional framework for achieving improved WASH targets, water coverage in hospitals is 85% or greater. To date, few LMICs have assessed or reported the availability of water in hospitals.
The World Health Assembly 68.15 underlined the importance of strengthening emergency and essential surgical care as a component of universal health coverage; in particular, it specifies the concern that inadequate investment in infrastructure, such as water availability, limits progress in improving delivery of surgical care.
The need to better defining the critical aspects of providing essential surgical care lead to the launch of the WHO Global Initiative for Emergency and Essential Surgical Care in 2005, an alliance of international organizations, civil and professional societies, nongovernmental organizations, and those representing disciplines of surgery, orthopedics, anesthesia, emergency medicine, and obstetrics.
Emergency and essential surgical care: Global Initiative for Emergency and Essential Surgical Care (GIEESC). WHO Web site. www.who.int/surgery. Accessed January 30, 2016.
Recognizing the availability of water at hospitals providing surgery as a limiting factor is important for this mission and surgeons operating in LMICs. Reliable water availability is essential for providing safe surgical care.
Water conservation in surgery: a comparison of two surgical scrub techniques demonstrating the amount of water saved using a ‘taps on/taps off’technique.
Water is required for washing instruments, steam sterilization, wound irrigation, and surgical hand scrub (SHS). SHS is an essential surgical safety measure
Emergency and essential surgical care: Global Initiative for Emergency and Essential Surgical Care (GIEESC). WHO Web site. www.who.int/surgery. Accessed January 30, 2016.
Water conservation in surgery: a comparison of two surgical scrub techniques demonstrating the amount of water saved using a ‘taps on/taps off’technique.
Although a number of surgical capacity assessments have reported a lack of equipment and supplies, water availability at hospitals remains poorly characterized.
These assessments provide a unique opportunity to examine the water availability at hospitals in LMICs. Thus, we aimed to systematically review the surgical capacity literature and describe the availability of water at hospitals in LMICs where they were performed. In addition, we used data from these assessments to model water availability at hospitals in LMICs more broadly. By doing so, the findings might identify fundamental limitations in surgical capacity and inform targeted capacity development strategies.
Material and methods
Systematic review
We designed a systematic search strategy to identify all published surgical capacity assessments that reported the water availability at hospitals in LMICs (see Supplementary material). The search strategy included terms for each of the following surgical care capacity assessments:
(i)
WHO's Tool for Situational Analysis to Assess Emergency and Essential Surgical Care
Validation of the World Health Organization tool for situational analysis to assess emergency and essential surgical care at district hospitals in Ghana.
Other terms were used to identify records that did not use the three tools aforementioned but assessed surgical care capacity, such as “surgical,” “surgery,” “capacity,” “assessment,” and “survey” (see Supplementary material). The World Bank World Development Report was used to define LMICs.
Low-income countries are those with a gross national income (GNI) per capita of $1045 or less in 2014; lower-middle-income countries are those with GNI per capita between $1046 and $4125; upper-middle-income countries are those with GNI per capita between $4126 and $12,735.
The titles and abstracts of retrieved records were screened for relevance, and the duplicates were removed. Two reviewers (S.S.C. and S.G.) screened all records; a third reviewer (A.L.K.) resolved disagreements. The remaining full-text reports and their reference lists were reviewed. Reports were included if they described the availability of water at one or more hospitals in a LMIC. If multiple reports from one country were found, the report with the most hospitals assessed was included.
Systematic review data analysis
A binary score was assigned to water availability for each hospital (i.e., water reliably available or not reliably available). For studies that did not provide data for each hospital, the countrywide prevalence of water availability was used. Most reports did not provide description of type or quality of the water source; thus, we were not able to stratify the reports by these important differences.
The percent of hospitals with water availability in each country were pooled using a fixed-effect model with Stata v13 (StataCorp, TX). Weights were applied based on the capacity assessment sample size (i.e., number of hospitals assessed), and a fixed continuity correction was added for countries that had water available at all hospitals assessed to allow the use of non-negative constants. Confidence intervals (CIs) were calculated with a binomial distribution. Estimates of heterogeneity were calculated from the inverse-variance fixed-effect model. Significant intragroup heterogeneity was identified (i.e., I2 was between 53%-93% with corresponding P values < 0.05).
To explore heterogeneity further, percentages were transformed using the Freeman–Tukey double arcsine method and pooled using a DerSimonian–Laird random-effects model for each income-level subgroup (i.e., low-income, lower-middle-income, and upper-middle-income countries).
Despite these attempts to minimize statistical heterogeneity, the I2 measure continued to be 50%-95%, which suggested substantial intragroup and intergroup heterogeneity. Thus, pooled estimates are not discussed. Instead, we present a narrative synthesis of water availability as recommended by the Cochrane Collaboration.
Ryan, R. (2015). Communication review group (2014). Heterogeneity and subgroup analyses in Cochrane Consumers and Communication review group reviews: planning the analysis at protocol stage.
Forest plots are presented to demonstrate both the range and variability of water availability between assessments.
Modeling water availability at hospitals in LMICs globally
National indicator data from the World Bank related to healthcare capacity and water availability were extracted for all LMICs, including the countries that had surgical capacity assessments.
These indicators included: gross domestic product (GDP), total healthcare expenditure, percent of rural population with improved access to water (rural), and percent of urban population with improved access to water (urban; Table 1).
Table 1Univariate variables used to build the predictive model of water availability.
Coefficient
Variable
Gross domestic product (GDP)
Total healthcare expenditure as % of GDP
Percent of rural population with improved access to water
Percent of urban population with improved access to water
We built a multivariable logistic regression model to predict of water availability at hospitals in LMICs more broadly. To do so, we examined bivariate relationships between national indicators and water availability at the hospitals reported by the capacity assessments. We then included all the predictors in the model simultaneously. Each of the national indicator covariates demonstrated evidence for a significant relationship with the percent water availability at hospitals (P < 0.05); thus, they were included in the multivariable model:
Crude odds ratios and multivariable odds ratios and their respective 95% CIs that describe the relationship between the percent of hospitals with water availability and national predictors were calculated from the logistic model.
Evaluating the goodness-of-fit and discriminating ability of the model
We assessed model accuracy (i.e., agreement between model-predicted percent of hospitals with water availability and observed percentages) using the Hosmer–Lemeshow goodness-of-fit test and calibration plots (i.e., graphs of predicted percentages versus the observed percentages of water availability). We grouped observations based on percentiles of ordered values of model-predicted probabilities such there was about the same number of observations in each group.
There was no evidence of poor model fit (Hosmer–Lemeshow chi-square test statistic P = 0.38). The calibration plots showed that observed and expected frequencies were very close to each other, which suggested that the logistic model fit the data well.
Next, we assessed the model's ability to discriminate (i.e., correctly classify hospitals with or without water availability from the capacity assessments) by calculating predictive performance indices (i.e., sensitivity and specificity values) for different model-predicted percentage cutoff values (Supplementary material). Using this estimates, we constructed a receiver-operating-characteristic curve to evaluate the accuracy of the model (Supplementary material). The area under the receiver-operating-characteristic curve was 0.73, which suggested that our model had relatively good capability to correctly classify hospitals as with or without water availability. Finally, we used the model to predict percent of hospitals in a LMIC with water availability and the respective 95% CIs.
Results
Search results
The systematic search returned 72 records; 50 records were not relevant to the study; no duplicates were identified. The remaining 22 full-text reports were reviewed. Of these, 19 reports described water availability at hospitals in a LMIC (Fig. 1).
Fig. 1Preferred Reporting Items for Systematic Reviews and Meta-Analyses flow diagram for the results form a systematic search for reports of water availability from surgical capacity assessments in low- and middle-income countries.
LMICs with at least one report of water availability spanned five continents: Africa (11 reports; 58% of reports); Asia (4; 21%); South America (2; 11%); North and Central America (1; 5%); and Oceania (1; 5%). Reports from three countries, Afghanistan, Somalia, and Ghana, did not include hospital-specific data regarding water availability. The number of hospitals per country ranged from nine in Guyana and the Solomon Islands to 48 in Tanzania (median number of hospitals 17; interquartile range [IQR] 14-28). In total, the reports described water availability at 430 hospitals in 19 LMICs.
Water availability reported by capacity assessments
Overall, 66% of hospitals assessed had water availability (283 of 430 hospitals). The median percent of hospitals with water availability from published reports was 61% (IQR 47.5%-82%). Only one assessment, from Bangladesh, reported that all of the hospitals assessed had water availability.
The percentage of hospitals with water access always available ranged from 22% in Malawi to 100% in Bangladesh (median 61%; IQR 48%-82%). In Malawi, 19 of 23 district hospitals and 2 of 4 central hospitals lacked a source of running water that was always available (Fig. 2).
Fig. 2Forest plot of water availability for countries included in assessment. (Color version of figure is available online.)
Countries with moderate to high GDP (i.e., GDP greater than 10 billion nominal USD) were more likely to have a higher percentage of hospitals with water availability than countries with GDP less than five billion nominal USD (Table 2). Total health expenditure was less strongly associated with the percent of hospitals with water availability, particularly in the multivariable model. While percent of rural population with improved access to water was associated with a higher percent of hospitals with water availability in both the bivariate and multivariable model, percent of urban population with improved access to water was not strongly associated with percent of hospitals with water availability in the multivariable model.
Table 2Predictors of the availability of water at the hospitals.
Estimated water availability from our model ranged from less than 20% in Sierra Leone, Liberia, and Togo to more than 90% in India, Malaysia, Thailand, Armenia, and Guinea (Table 3). In general, the sub-Saharan African region had the lowest estimated water availability and South and Southeast Asia regions had the highest estimated availability (Fig. 3).
Table 3Estimated percent of hospitals with a reliable water source in low- and middle-income countries.
Country
Lower estimate
Estimate
Upper estimate
East Asia and Pacific
Kiribati
3.5
14.4
43.5
Solomon Islands
20.4
34.7
52.3
Timor-Leste
16.9
41.9
71.9
Samoa
28.3
49.3
70.5
Mongolia
38.9
52.6
66
Fiji
35.4
61.6
82.5
Tonga
39.5
66.2
85.4
Cambodia
62
76
86
Laos
59
80.7
92.4
China
71.6
83.7
91.3
Myanmar
71.8
85.1
92.8
Vietnam
73.1
85.4
92.7
Indonesia
74.7
86.9
93.7
Philippines
78.2
89.4
95.2
Malaysia
81.2
91.2
96.1
Thailand
81.6
91.7
96.4
Europe and Central Asia
Montenegro
30
50.7
71.1
Azerbaijan
67
77.3
85.1
Serbia
64.7
80.4
90.2
Albania
68.9
83.2
91.7
Ukraine
70.6
84.6
92.6
Macedonia
71.2
85.8
93.6
Bulgaria
73.6
86.1
93.3
Belarus
73.8
86.2
93.3
Bosnia and Herzegovina
73.7
86.4
93.5
Georgia
73.9
86.5
93.6
Romania
73.9
86.5
93.6
Turkey
73.9
86.5
93.6
Tajikistan
71
88.5
96.1
Moldova
83.5
88.7
92.4
Kazakhstan
78.5
89.6
95.3
Armenia
79.1
92.2
97.4
Kyrgyzstan
76
92.9
98.2
Latin America and the Caribbean
Grenada
28.5
47.9
67.9
St. Lucia
28.9
48.3
68.4
Guyana
28.6
49.1
69.9
Haiti
42
72
90.1
Peru
62.6
72.2
80.1
Dominican Republic
60.5
75
85.5
Nicaragua
50.2
75.3
90.2
Ecuador
65.8
75.8
83.7
Colombia
65.6
76.6
84.9
Bolivia
66.5
77.3
85.3
Honduras
69.8
80.7
88.3
El Salvador
70.5
81.6
89.2
Guatemala
71
82
89.5
Panama
71.1
82.4
89.9
Brazil
71.5
82.7
90
Mexico
71.4
83.4
91
Costa Rica
72.6
84.1
91.3
Paraguay
73.3
85.1
92.2
Middle East and North Africa
Djibouti
15.8
30.6
50.8
Iraq
63.4
73.7
81.9
Morocco
59.2
73.8
84.5
Algeria
59.2
74.5
85.5
Iran
71.6
83.5
91.1
Jordan
71.7
83.6
91.1
Lebanon
73.2
86
93.2
Egypt
73.8
86.2
93.3
South Asia
Maldives
20.1
39.7
63.2
Afghanistan
44.6
53.4
61.9
Bhutan
39.9
66.5
85.6
Bhutan
39.9
66.5
85.6
Nepal
66.4
81.1
90.3
Bangladesh
73.1
86.8
94.1
South Asia
Pakistan
78.2
89.3
95.1
India
80.2
90.6
95.8
Sri Lanka
81.3
91.3
96.2
Sub-Saharan Africa
Sierra Leone
3.1
13
40.9
Liberia
5.7
18.5
46
Togo
7.9
19.3
40.2
Eritrea
9.9
26.7
54.6
Lesotho
10.2
26.7
53.9
Guinea Bissau
13.7
29.2
51.6
Swaziland
17.4
30.7
48.3
Burundi
18.5
31.8
49
Central African Republic
13.6
35.5
65.8
Gambia
23
39
57.8
Comoros
23.1
40.7
61.1
Malawi
25
42.5
62
Sao Tome and Principe
28
46.9
66.7
Mozambique
35.6
49.1
62.8
Tanzania
43
52
60.7
Cabo Verde
30.7
56.4
79
Angola
30.6
57
79.9
Zambia
49.4
60.1
69.9
D.R.C.
33.7
61.9
83.8
Ethiopia
44.7
62.8
77.9
Chad
35.8
63.3
84.2
Madagascar
27.4
63.6
89
Cameroon
48.4
65.9
79.9
South Sudan
43.1
66.8
84.3
South Sudan
43.1
66.8
84.3
Mali
46
71.6
88.2
Sub-Saharan Africa
Cote d'Ivoire
62.6
72.8
81
Republic of Congo
28.9
73.1
94.8
Nigeria
56.9
74.5
86.6
Kenya
56.9
74.6
86.8
Uganda
66.5
76.9
84.8
Burkina Faso
56.7
77.2
89.7
Ghana
67.3
78.9
87.1
South Africa
69.5
80.6
88.4
Namibia
63.4
80.8
91
Senegal
67.4
83
92
Rwanda
77.9
83.2
87.5
Mauritania
56.8
83.8
95.3
Botswana
72.5
84.1
91.4
Gabon
67
84.2
93.3
Niger
65.3
85.7
95
Mauritius
79
92.1
97.3
Benin
84.4
93.2
97.2
Guinea
86.8
93.6
97.1
Data unavailable for: American Samoa, Belize, Cuba, Dominica, Jamaica, North Korea, Kosovo, Libya, Marshal Islands, Micronesia, Palau, Papua New Guinea, Somalia, Sudan, Suriname, Syria, Tunisia, Turkmenistan, Tuvalu, Uzbekistan, Palestine, Yemen, and Zimbabwe.
Fig. 3Map of estimated percent of hospitals with a reliable water source in low- and middle-income countries. (Color version of figure is available online.)
This study aimed to assess water availability at hospitals in LMICs. Water is an essential infrastructure resource for providing safe surgical care. Given that only 66% of hospitals that underwent surgical capacity assessment had water availability, this resource deficiency likely represents a common and significant barrier to the provision of surgical care. Our model highlights this potential deficiency in many LMICs. Thus, specific attention to hospital infrastructure development to facilitate surgical capacity improvements is urgently needed.
Other than the reports described, few data are available to put our results into context. Reports of other infrastructure deficiencies (e.g., electricity) at hospitals in many LMICs suggest that improvements to these essential resources are required. Similarly, there are no data regarding the effect of improving infrastructure on surgical care output. However, water availability within hospitals is essential to not just providing surgical care but also preventing disease. In addition to facilitating safe surgery (e.g., SHS, steam sterilization, and wound irrigation), waterborne nosocomial infections in LMICs can lead to increased patient morbidity, mortality, and hospitalization costs.
Given these broader benefits, water availability should be prioritized for improving access to safe health care and surgical care provision.
The first step to addressing the challenge of water availability in hospitals in LMICs is to set national policies with appropriate funding and support. Examples of successful implementation of water availability strategies have been reported from Laos and Bangladesh.
In 25 hospitals in Laos, the national government provided guidance on environmental standards, integration of standards into national programs, and support for capacity building and personnel training.
The success is reflected in our model, which predicts water availability in hospitals in Laos of 81%, well above the median percentage for LMICs in general (66%).
Another success story is Bangladesh, which has provided water availability to all hospitals countrywide. This initiative was created given the high incidence of diarrheal disease, which remains a major cause of death among all ages in the country.
Thus, future capacity assessments might consider assessing the temporal water availability, water quality, and type of water infrastructure so that specific recommendations can be made.
An unexpected result of the model was that it suggests that some lower GDP countries were more likely to have hospitals with water availability than relatively high GDP countries. While this may be due to uncontrolled confounding among unmeasured variables, it may also be the result of WASH programs that are under funded and/or under resourced in spite of relative national wealth.
The model might be useful for identifying countries that may benefit from improving hospital water availability. First, the model needs to be validated by assessments of water availability at hospitals in several of the countries reported. Should it be found to be valid, national governments and international organizations might use the data from the model to advocate and plan for improving hospitals' access to reliable and clean water. Given the dearth of data regarding the impact of water availability on healthcare outcomes in LMICs, monitoring and evaluation exercises could also prove beneficial as capacity improvement activities commence. The model may be by countries to better understand their level of water infrastructure, as it relates to surgical capacity. For example, bordering countries may be at opposite ends of water availability; notably Burundi has a predicted availability of 31.8% whereas neighboring Rwanda has a predicted availability of 83.2%. Burundi may possibly gain from the strategies used by Rwanda to build its water and surgical capacity infrastructure.
This study has a number of limitations. While our search identified assessments from 19 countries, the assessments varied in the number of hospitals visited and the reporting of quality. Thus, we were unable to describe water availability with high accuracy or in greater detail. Second, assessments rarely reported the type of water source or its temporal availability; thus, we were forced to use binary classification of availability: availability or unavailable. This classification might have misclassified hospitals with intermittent water availability making our estimates overly conservative. Third, our model suggested that hospitals in some countries with relatively high GDP were less likely to have water available than hospitals in countries with lower GDP. In addition, we were not able to make a reasonable comparison between water availability at hospitals and surgical care. At last, robust data to validate our model are not available. However, despite these limitations, the findings suggest that water availability at hospitals in LMICs is generally insufficient and requires at least systematic assessment, if not significant improvement.
Conclusions
The availability of water at hospitals in LMICs is likely insufficient globally and may be hindering the provision of safe surgical care. There is wide variation in availability between and within regions that requires specific attention to determine the best courses of action given differences in current availability, demand, infrastructure, and natural resources. The model presented offers a useful starting point for national and international organizations to consider when planning healthcare and surgical capacity improvement initiatives. Securing improved water sources at the household level has been one of the greatest successes of the United Nations Millennium Development Goals; we must extend this success to hospitals. This will require concerted effort by LMICs and implementing partners to assess the current situation, develop appropriate policies, set national targets, and building capacity in this fundamental healthcare resource.
Acknowledgment
This study was funded in part by grant R25-TW009345 from the Fogarty International Center, US National Institutes of Health. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
Authors' contributions: All authors had substantial contribution to the manuscript. S.S.C. was responsible for data abstraction, analysis, and manuscript writing. B.T.S. contributed to the manuscript writing and data analysis. F.M.O. contributed to modeling analysis and manuscript writing. S.G., E.B.H., and A.L.K. contributed through project guidance, data analysis, and manuscript writing.
Disclosure
The authors have no real or potential conflict(s) of interest to disclose.
Emergency and essential surgical care: Global Initiative for Emergency and Essential Surgical Care (GIEESC). WHO Web site. www.who.int/surgery. Accessed January 30, 2016.
Water conservation in surgery: a comparison of two surgical scrub techniques demonstrating the amount of water saved using a ‘taps on/taps off’technique.
Validation of the World Health Organization tool for situational analysis to assess emergency and essential surgical care at district hospitals in Ghana.
Ryan, R. (2015). Communication review group (2014). Heterogeneity and subgroup analyses in Cochrane Consumers and Communication review group reviews: planning the analysis at protocol stage.
30161-5&id=gr1.jpg){kind=link}
30161-5&id=gr2.jpg){kind=link}
30161-5&id=gr3.jpg){kind=link}