Impact of rotavirus vaccine on diarrheal hospitalization and outpatient consultations in the Philippines: First evidence from a middle-income Asian country.

BACKGROUND: Monovalent rotavirus (RV) vaccine was introduced in the Philippines in a phased manner beginning in 2012. To assess the impact of RV vaccine, we conducted a retrospective review of diarrheal admissions in two hospitals.
METHODS: Records of physician-diagnosed diarrheal admissions were reviewed in D.O. Plaza Hospital (DOPH) from 2009 to 2016 in Agusan del Sur where RV vaccine was introduced in the immunization program; and in Cotabato Regional Medical Center (CRMC) from 2011 to 2016 in a region where the vaccine was not introduced. Reports from consultations in public health clinics in Agusan Del Sur and RV vaccine coverage were obtained.
RESULTS: All-cause diarrheal admissions among children <5 years old in DOPH declined from 2013 to 2016 following RV vaccine introduction in 2012. Using the 2009-2011 mean number of hospitalizations as baseline (X‾ = 1,141), the reductions were 28% (n = 821), 56% (n = 507), 63% (n = 417) and 59% (n = 466) in 2013, 2014, 2015 and 2016, respectively. In comparison, no substantial declines in diarrheal hospitalizations were seen in CRMC from 2011 to 2016. A declining trend was also seen in outpatient consultations in Agusan del Sur following RV vaccine introduction with declines of 27% (n = 2,333), 33% (n = 2,143), 45% (n = 1,764) and 67% (n = 1,059) in 2013, 2014, 2015 and 2016. From September 2012 to December 2016, the 1 and 2-dose RV vaccine coverage gradually increased from 5% and 4% in 2012 to 92% and 88% in 2015, but decreased in 2016 to 53% and 52%, respectively. DISCUSSION: RV vaccine introduction was associated with a substantial decline in diarrheal hospitalizations and outpatient consultations for diarrhea in Agusan del Sur, Philippines.

Introduction

Globally in 2015, diarrhea ranked as the fourth leading cause of death among children under 5 years of age [1] and the fourth leading cause of disability in this age group [2], resulting in 499,000 deaths [3]. In the Philippines in 2013, diarrhea accounted for 5169 estimated deaths among children under 5 years, comprising 7.3% of all deaths in this age group [4].

Rotavirus (RV) was the most common cause of diarrheal death globally in 2015, accounting for an estimated 146,000 deaths among children under 5 years old [3]. In the Philippines, it was estimated that in 2013, RV was responsible for 3.7% of all deaths among children under 5 years or 2599 deaths [5].

Because of the substantial burden of RV diarrhea, the World Health Organization (WHO) has recommended inclusion of RV vaccine in national immunization programmes of all countries globally since 2009 [6]. In July 2012, the Philippines added RV vaccine to its immunization program and made the monovalent RV vaccine (Rotarix®, GlaxoSmithKline) available free-of-charge to infants of families belonging to the lowest economic quintile. However, prior to nationwide introduction, policymakers requested further evidence on the effect of the introduction of RV vaccine in the Philippines. The Department of Health (DOH) identified the province of Agusan del Sur to conduct an effectiveness study due to the high number of diarrhea hospitalizations in young children, the capacity for surveillance and the relatively lower than average socioeconomic status of households in the province.

In Agusan del Sur, RV vaccine became available to the poorest quintile in September 2012, and in January 2013, availability was expanded to all age-eligible children in two municipalities, San Francisco and Prosperidad, and in July 2014 to the whole province. The two-dose course of RV vaccine is given to infants aged 6 weeks and 10 weeks. We assessed the impact of RV vaccine introduction in the province of Agusan del Sur by comparing diarrheal hospitalizations and deaths in the provincial public hospital and outpatient consultations for diarrhea in public health centers, before and after the introduction of RV vaccine. We also compared diarrheal hospitalizations in the Agusan del Sur provincial hospital with those in Cotabato Regional Medical Center (CRMC), where RV vaccine was not yet introduced in public health clinics.

Methods

Population and setting

Agusan del Sur is a land-locked province in the northern part of the island of Mindanao and has 13 municipalities and one city (Fig. 1). It is one of the most impoverished provinces in the Philippines with 58% of the population living below the poverty line [7]. D. O. Plaza Hospital (DOPH), located in the municipality of Prosperidad, is the biggest public, tertiary hospital, catering to the population of Agusan del Sur along with portions of another province, Surigao del Sur. A review of DOPH data from 2007 to 2012 indicate that approximately 70% of diarrhea cases seen in DOPH were from the municipalities of Prosperidad and San Francisco in Agusan del Sur.

Fig. 1

Study sites: Location of the island of Mindanao in the Philippines (A); Location of Agusan del Sur, served by D.O. Plaza Hospital and Cotabato City and North Cotabato, served by Cotabato Regional Medical Center (B).

CRMC in Cotabato City is a tertiary training and teaching hospital serving Cotabato City and North Cotabato. Cotabato City is located on the central and western part of the island of Mindanao and North Cotabato is adjacent to the city (Fig. 1). In 2012, the poverty levels of North Cotabato and Cotabato City were 42% and 44% [8]. CRMC has been a sentinel site for RV surveillance since June 2013, and registered the second highest number of cases in the surveillance with 251 cases in 2013 (second to DOPH).

Eligibility and data collection

Patient records were eligible for inclusion if the patients: (1) presented with diarrhea and were diagnosed as acute gastroenteritis or acute diarrhea by physicians, (2) were <5 years of age at the time of diagnosis, and (3) were residents of Agusan del Sur admitted between January 1, 2009 and December 31, 2016 (for DOPH) or were residents of Cotabato City or North Cotabato province and admitted between January 1, 2011 and December 31, 2016 (for CRMC). We excluded patients who had nosocomial diarrhea and those with chronic diarrhea. A uniform data extraction form was used for identified cases and entered in an electronic database (Microsoft® Excel for Mac 2016 ver. 15). The following information was collected: demographic profile (age, sex, address/location) and clinical profile for hospitalized cases (date of admission or outcome, degree of dehydration, length of hospitalization).

In addition, for Agusan del Sur, reports from the Field Health Service Information System (FHSIS) of the Epidemiology Bureau of the DOH were obtained. FHSIS collates data on national programs of the DOH (e.g., child care, immunization, family planning use, etc.) in public clinics. Data from public community clinics called Barangay (community) Health Stations and Rural Health Units across the country are aggregated per clinic and submitted to FHSIS regularly. No individual-level information from the clinics was available. From this database, we obtained the number of children with diarrhea given oral rehydration therapy (ORT/ORS) with or without zinc and immunization coverage for the province as a whole, and disaggregated by municipality.

Data analysis

Trends in diarrheal hospitalizations in DOPH

Demographic and clinical characteristics of hospitalized all-cause diarrhea admissions were tabulated. For DOPH, the number of cases during the pre-vaccine introduction (baseline) period (from January 2009 to December 2011) was compared to the number during the post-introduction period (from January 2013 to December 2016). We excluded information from 2012 as this was the transition year when vaccine was being introduced and the vaccine was not universally available to all children.

In-hospital diarrheal deaths and diarrheal hospitalizations were tabulated by age group. Absolute numbers and proportions were compared among age groups. Statistical significance was assessed using chi-square with significance set at p < 0.05.

Trends in outpatient consultations in Agusan del Sur

Diarrheal consultations managed with ORS and zinc in public health clinics were tabulated annually. The mean number of diarrheal consultations from 2009 to 2011 was calculated and used as a baseline. As with hospitalizations, data from 2012 were excluded.

Comparison of diarrhea hospitalization trends between DOPH and CRMC

Data from 2009 to 2010 were unavailable from CRMC. We compared data from DOPH and CRMC for 2011–2016.

Ethics

The study was approved by the University of the Philippines Manila Review Ethics Board (UPMREB 2015-L31-01). This evaluation involved analysis of existing, publicly available datasets (reports, registries and census) and collection of de-identified data from patient hospital records, with no patient interaction. No informed consent was required.

Results

Trends in diarrheal hospitalizations in DOPH

From January 1, 2009 to December 31, 2016, there were 6544 cases of diarrhea among children under 5 years of age who were admitted in DOPH, including 6039 discharges and 94 deaths (Table 1 and Fig. 2), while 67 transferred to another hospital and 344 left against medical advice. Almost one-third (31%) of the diarrhea admissions occurred in children aged 6–11 months, while 31% of the 94 in-hospital diarrhea deaths were in the 0–5 month age group. Sixty-one (65%) of the diarrhea deaths occurred in children who were assessed to have severe dehydration on admission. During the pre-vaccination years 2009–2011, diarrheal hospitalizations averaged 1141 per year (Table 2). After RV vaccine introduction, hospitalizations decreased by 28% (n = 821), 56% (n = 507), 63% (n = 417) and 59% (n = 466) in 2013, 2014, 2015 and 2016 respectively compared to pre-vaccination years (Fig. 3). The largest declines were seen in the 6–11 month age group (Fig. 4). The highest number of hospitalizations and deaths occurred in 2011 with 1361 cases and 33 deaths. Deaths due to diarrhea, by year, were 12 (2009), 7 (2010), 33 (2011), 9 (2012), 9 (2013), 6 (2014), 9 (2015) and 9 (2016). Excluding 2011 and using 2009–2010 as baseline years ( = 1031), the declines in diarrheal admissions at DOPH were 20%, 51%, 60% and 55% in 2013, 2014, 2015 and 2016, respectively.

Table 1

Characteristics of children <5 year of age admitted for diarrheal diseases in D. O. Plaza Memorial Hospital, January 2009–December 2016.

Characteristic	All diarrheal admissions N = 6544	Discharges N = 6039	Deaths N = 94	P valuea
Age (in months), mean ± s.d.	15.7 ± 8.8	15.9 ± 8.8	13.8 ± 9.7	0.103
Age group, n				0.0001
0–2 mos	322 (4.9)	289 (4.8)	12 (12.8)
3–5 mos	575 (8.5)	511 (8.5)	17 (18.1)
6–8 mos	1007 (15.3)	925 (15.3)	13 (13.8)
9–11 mos	1040 (15.8)	955 (15.8)	9 (9.6)
12–17 mos	1508 (23.3)	1408 (23.3)	15 (16.0)
18–23 mos	834 (12.8)	775 (12.8)	10 (10.6)
24–59 mos	1258 (19.5)	1176 (19.5)	18 (19.1)
Level of dehydration, n				<0.00001
No dehydration	42 (0.6)	38 (0.6)	0 (0)
Some dehydration	6373 (97.4)	5948 (98.5)	33 (35.1)
Severe dehydration	129 (2.0)	53 (0.9)	61 (64.9)

p value comparing characteristics among discharges and deaths.

Fig. 2

Monthly trends in the number of all-cause diarrheal hospitalizations and deaths among children <5 years of age in D.O. Plaza Hospital, Agusan del Sur, 2009–2016.

Table 2

Percent reduction in all-cause diarrheal hospitalization among children <5 years of age, by age group and year compared to the pre-vaccine baseline.

	Baseline	2013		2014		2015		2016
	(2009–2011)
	N	N	Reduction% (95% CI)	N	Reduction% (95% CI)	N	Reduction% (95% CI)	N	Reduction% (95% CI)
0–5 mos	141	109	23% (15, 31)	56	60% (47, 73)	79	44% (33, 55)	85	40% (30, 50)
6–11 mos	378	253	33% (27, 39)	142	62% (54, 70)	117	69% (61, 77)	124	67% (59, 75)
12–23 mos	405	314	22% (17, 27)	169	58% (51, 65)	151	63% (55, 71)	156	61% (53, 69)
24–59 mos	217	145	33% (25, 41)	140	36% (28, 44)	70	68% (57, 79)	101	54% (44, 64)
Total	1141	821	28% (25, 31)	507	56% (52, 60)	417	63% (58, 68)	466	59% (55, 64)

Fig. 3

Number of all-cause diarrheal hospitalizations among children <5 years of age in D.O. Plaza Hospital (DOPH) from 2009 to 2016 and Cotabato Regional Medical Center (CRMC) from 2011 to 2016 (including linear trending).

Fig. 4

Number of all-cause diarrheal hospitalizations among children <5 years of age in Democrito O. Plaza Hospital, 2009–2016 (A) and Cotabato Regional Medical Center, 2011–2016 (B), by age-group.

Comparison of diarrheal hospitalizations between CRMC and DOPH

In comparison, in CRMC, there were 4944 diarrheal hospitalizations from 2011 to 2016, and among these there were 3424 discharges (69%), 102 deaths (2%), 1409 that left against medical advice (29%), and 9 transfers (0.1%) (Fig. 3). During this period, the number of cases remained stable, with an average of 824 per year and a flat linear trend except in 2015 when there were 1150 diarrheal admissions. The age distribution of cases in CRMC was significantly different from that in DOPH (p < 0.001). In CRMC, the majority of cases were aged 6–11 months (n = 1386, 28%) or 12–23 months (n = 1560, 31.6%) (Fig. 4). The outcomes of the diarrheal cases in CRMC were significantly different (p-value < 0.00001) from the outcomes of diarrheal cases in DOPH during the years 2011–2016. Diarrheal deaths in CRMC were 20, 19, 24, 14, 16 and 9 in 2011, 2012, 2013, 2014, 2015 and 2016, respectively.

Trends in diarrheal consults in Agusan del Sur

In Agusan del Sur, there were 19,607 recorded diarrheal consultations requiring ORS among children <5 years of age in all community clinics from January 2009 to December 2016 (Fig. 5). Similar to diarrheal hospitalizations, the number of consultations (Fig. 5) was relatively stable from 2009 to 2010 and increased in 2011 ( = 3210). Compared to baseline years of 2009–2011, after RV vaccine introduction there were declines of 27% (n = 2333), 33% (n = 2143), 45% (n = 1764) and 67% (n = 1059) in 2013, 2014, 2015 and 2016 respectively. The highest number of diarrhea consultations occurred in 2011 with 3817 cases. Excluding 2011 and using 2009–2010 as baseline years ( = 2906), diarrhea consults declined by 20%, 26%, 39% and 64% in 2013, 2014, 2015 and 2016, respectively, in the public health clinics.

Fig. 5

Number of diarrheal hospitalizations and diarrheal outpatient consultations among children <5 years age versus rotavirus vaccine (RV) coverage among children <1 year of age in the Province of Agusan del Sur, 2009–2016.

Immunization coverage

RV vaccine coverage among children under 1 year of age increased from 5% for RV vaccine dose 1 and 4% for RV vaccine dose 2, respectively, in 2012, and to 92% and 88%, respectively, in 2015 (Fig. 5). Coverage decreased in 2016 to 53% for dose 1 and 52% for dose 2. Vaccine stock-outs occurred from October 2013 to June 2014, and from June to December 2016 (Fig. 2).

Discussion

Compared to the 2009–2011 baseline period, the number of all-cause diarrhea hospitalizations and outpatient consultations in public clinics among children <5 years of age in Agusan del Sur declined in 2013–2016 and was more than 50% lower than the baseline for every year after 2013. This decline was temporally associated with the introduction of RV vaccine into the vaccination schedule for infants aged 6 weeks and 10 weeks. In comparison, in Cotabato City and North Cotabato, where no RV vaccine was introduced, no decline in all-cause diarrhea hospitalizations was seen among children <5 years of age.

The consistent and progressively declining diarrheal hospitalizations in DOPH and decline in outpatient consults for diarrhea coincident with increasing vaccine coverage in Agusan del Sur and the absence of decline in diarrheal hospitalizations in CRMC where RV vaccine was not available in the public health program and the lack of change in breastfeeding coverage and access to water and sanitation throughout the duration of the analysis provide support to the positive impact of RV vaccine on diarrheal disease. In addition, the biggest declines in diarrheal cases were among children 6–24 months of age which is the peak age group for RV disease. Larger declines in diarrhea hospitalizations and in-hospital deaths occurred over time as the scope of the vaccination program expanded from selected introduction in 2012 to regional introduction in 2014 and as RV vaccine coverage increased. Our findings are further supported by data from the national RV surveillance that began in 2012 [9] in a phased manner. In DOPH and CRMC, surveillance started in 2013. A declining trend in the proportion of RV positive cases was seen in DOPH (47%, 41%, 31%, 27%, annually from 2013 to 2016) whereas no trend can be identified in CRMC (29%, 36%, 33%, 37% annually) during the same period.

RV is responsible for a large proportion of hospitalized diarrhea cases [10] and RV is one of the leading causes of moderate to severe diarrhea in children under 5 years, particularly in infants [11]. In the Philippines, prospective studies have shown that RV accounted for 30–40% of all-cause diarrhea cases [12], [13], [14], [15], [16], [17], [18], [19], [20], [21] and initial results from surveillance showed that RV was responsible for approximately 40% of hospitalized diarrhea cases among children under 5 years of age [9]. Reductions in all-cause diarrhea hospitalizations and deaths have been seen following RV introduction in various countries [22], [23], [24]. In our study, reductions of up to 63% were seen in Agusan del Sur while no changes were seen in CRMC. Although the larger declines in hospitalizations and diarrhea cases could reflect overall improvements in the standard of living as the economy improved, by 2015 Agusan del Sur, North Cotabato and Cotabato City all had similar economic standing [8].

Our results should be interpreted with caution. This is an ecologic study subject to ecologic bias and causality cannot be determined on an individual basis [25]. We included all hospitalizations and out-patient consultations due to diarrhea, not just due to RV alone. Other etiologic agents may cause an increase in the number of diarrhea cases and affect the interpretation of the results. In 2011, the Philippine Integrated Disease Surveillance and Response reported a cholera outbreak with cases coming from half of the municipalities in Agusan del Sur and this outbreak may have positively skewed the pre-vaccine introduction case numbers. However, cholera cases have been reported in Agusan del Sur and North Cotabato annually from 2008 to 2013 [26]. Furthermore, cholera is fairly uncommon in children under 2 years of age in less endemic countries like the Philippines. Declines in diarrheal hospitalizations and outpatient visits were still observed even after data from 2011, the peak year for cholera cases in Agusan del Sur, was excluded from the pre-vaccine baseline.

Some limitations should be noted. First, RV testing data were not locally available. Although DOPH and CRMC are sentinel surveillance hospitals for RV, specimens are tested at the national laboratory, the Research Institute for Tropical Medicine (RITM) in Metro Manila. Since surveillance was initiated only in 2012, data are not available for RV cases during the pre-vaccine introduction years and a direct comparison of RV case numbers before and after vaccine introduction is not possible. Second, vaccine stock-outs during two periods after introduction resulted in declines in vaccine coverage. Hence, we were unable to present two complete years of consistent vaccination coverage. However, the reduction in diarrheal hospitalizations plateaued in 2016, coincident with the decline in RV vaccine coverage that year.

Third, we were unable to get reliable information on deaths among children for the whole province. Although we tried to obtain information from the Agusan del Sur Civil Registry, deaths particularly in children, were not consistently recorded. According to the Philippines Statistics Authority it is estimated that in 2010, although 94% of births were recorded, only 66% of all deaths were captured in the registry. In remote mountainous areas such as in Agusan and North Cotabato, registration coverage may even be lower [27]. For this analysis, we only included the reported in-hospital deaths at DOPH and CRMC. A higher proportion of children hospitalized for diarrhea in CRMC than in DOPH went home against medical advice and it is unclear what their outcomes were. It may be that the actual number of deaths among children who presented in CRMC were higher than we report. Since CRMC is the only referral center in the area, it is unlikely that children were transferred to another public hospital.

Fourth, RV vaccines are available in the private market in the Philippines and this may have affected our results. However, the EPI estimates that ∼10% of the country’s children obtain their vaccines from private pediatricians, which are mostly concentrated in large urban areas. Considering the socio-demographic characteristics of the areas included in our study, it is likely that the proportion that received RV vaccines in the private sector is even lower than 10% and thus, no substantial effects would be seen in diarrheal disease trends in these areas.

Lastly, we conducted the records review in public hospitals and we were unable to include any private hospital. Similarly, FHSIS data only covered public community clinics. However, since the vaccine was provided in the public sector, vaccinated children are likely to have sought consultation in public clinics and hospitals.

Prior to nationwide introduction, policymakers requested evidence on the effect of RV vaccine introduction in the Philippines. The Philippines has never been Gavi-eligible and as in other lower middle-income countries, resources for vaccine procurement come from very limited budgets. For diarrheal diseases, policymakers considered that increasing rates of exclusive breastfeeding in the first 6 months of life and access to improved water and sanitation may be viable alternatives to RV vaccination to reduce diarrheal case numbers. These interventions are proven effective for reducing the risk of diarrheal disease and death, but even in settings with high standards of hygiene, RV infection is nearly universal in the absence of vaccination. WHO recommends the use of rotavirus vaccine as part of a comprehensive strategy to control diarrheal diseases through scaling up of both preventive and treatment services [28].

We present the first results on the impact of RV vaccine when given at 6 and 10 weeks of age in a public health program in a lower middle-income country in Asia. Because of the differential protection afforded by the vaccine in various economic settings, data from this setting provide important information for policymakers allocating scarce resources. The substantial decline in all-cause diarrheal hospitalizations following RV vaccine introduction supports implementation of RV vaccination to prevent diarrheal morbidity in this setting.

Funding

This work was supported in part by the World Health Organization.

Declarations of conflicts of interest

None