EVALUATION OF A MATERNAL AND CHILD ELECTRONIC

HEALTH RECORD IN A DEVELOPING COUNTRY

Preliminary Results from a Field Pilot

Martins Mupueleque

1,3

, Juliano Gaspar

1,2

, Ricardo Cruz-Correia

1,2

and Altamiro Costa-Pereira

1,2

CIDES – Department of Health Information and Decision Sciences, Porto, Portugal

CINTESIS – Center for Research in Health Technologies and Information Systems, Faculty of Medicine

University of Porto, Porto, Portugal

Faculty of Health Sciences, University of Lúrio, Nampula, Mozambique

Keywords: Maternal and child health, Childbirth, Information systems, Electronic health record (EHR), Data quality,

Mozambique, Africa.

Abstract: Introduction: The poor quality of health indicators data such as maternal and perinatal mortality prevents an

adequate planning and evaluation of public health interventions.

Objective: To evaluate the viability and usability of a prototype of a maternal and child electronic health

record, in Northern Mozambique.

Methods: The prototype was based on existing paper forms and the system was assessed at a maternity ard

in a healthcare center of Nampula. The usability was evaluated using in loco observations, interviews and

self-administered questionnaires to health professionals, after a 6 hours training period.

Results: During the evaluation period, 205 clinical forms were registered, corresponding to 58% of all

childbirths. From a total of 19 health professionals, 9 nurses answered the questionnaire. The prototype was

well accepted among these respondents but the recording time was higher than in its paper version (12 vs. 5

minutes per record). Data quality problems were identified. The absence of a unique citizen identifier was a

major obstacle regarding the mother-child data linkage.

Conclusion: Although this system may contribute to the improvement of the quality of the healthcare

provided to women and children in Mozambique, there are still many technical and organizational

challenges to be overcome.

1 INTRODUCTION

Approximately 11 women die every day in

Mozambique, adding up to 3.900 deaths in 2009,

from complications related with pregnancy and

childbirth. For every 1.000 live births, about 48

children die in the first 28 days of life every year,

mainly from complications occurring during

pregnancy and childbirth, including inadequate

home practices (2009a).

Nampula is a province in the Northern part of

Mozambique, covering an area of 78.197 km

, with

almost 4 million inhabitants, mainly farmers with

little education and strong ties to a traditional

medical system.

The inadequate infrastructure of Mozambique,

such as communication and transportation networks,

access roads, electricity, data processing equipment

and Internet, complicates the accessibility to

information, both paper-based and electronic (2007;

2008; 2009b).

During the research and the analysis of reports

regarding the Mozambican context, the authors

faced two main problems: the first one is connected

with the tremendous lack of data concerning

different health indicators, and the second one

concerns the quality of available information (OMS,

2010). The number of premature deaths (mortality)

is a strong indicator of this situation, as well as the

high levels of suffering (morbidity) and incapacity

(NEPAD, 2008); (Elizabeth et al., 2007).

1.1 Electronic Health Records

Over time, paper-based systems have become

increasingly systematized and gone through several

256

Mupueleque M., Gaspar J., Cruz-Correia R. and Costa-Pereira A..

EVALUATION OF A MATERNAL AND CHILD ELECTRONIC HEALTH RECORD IN A DEVELOPING COUNTRY - Preliminary Results from a Field Pilot.

DOI: 10.5220/0003788902560262

In Proceedings of the International Conference on Health Informatics (HEALTHINF-2012), pages 256-262

ISBN: 978-989-8425-88-1

 2012 SCITEPRESS (Science and Technology Publications, Lda.)

transformations. Paper-based medical records render

the communication between healthcare providers

extremely difficult, especially in developing

countries (Kalogriopoulos et al., 2009); additionally,

some physicians resist new technologies and prefer

the paper-based recording method (Hayrinen et al.,

2008).

Indisputably, individual patient data, collected

and accessed at the point of care, using an Electronic

Health Record (EHR), can assist clinical

management (Gladwin et al., 2002); (Zandieh et al.,

2008). Physicians can easily access previous records

and receive alerts to potential problems, such as drug

interactions (Fraser et al., 2005); (Dumont et al.,

2009); (Slagle, 1999).

Several African countries still face

organizational difficulties such as lack of data

quality; defective data analysis; timely data delivery

problems and data re-use problems by health

professionals. Moreover, the data collection

mechanism in Mozambican National Health Service

healthcare units is based in manual survey forms,

like annual questionnaires (2007; 2005; Allotey and

Reidpath, 2000).

Opinions regarding the security and

confidentiality of medical data differ across

developing countries. In some locations, the uses of

electronic databases are viewed with a good deal of

suspicion (Fraser et al., 2005); (Zandieh et al.,

2008). Patients can face serious risks in their

communities if, for instance, their HIV/AIDS status,

or other sensitive medical information, is uncovered.

These obstacles emanate from a poor medical

culture, where much of the information is rarely

searched for, usually unavailable and frequently

uncollected (Berkowicz et al., 1997).

1.2 Motivation

Approximately 536.000 deaths worldwide relate to

pregnancy and childbirth every year, and 99% of

these deaths take place in developing countries, 86%

of them in Sub-Saharan Africa (NEPAD, 2008;

2005). However, these estimates are collected from

different systems and many of them are unreliable

(Allotey and Reidpath, 2000). In order to enhance

the accuracy of these estimates we need to improve

the collection of data from pregnant women and

health care professionals, using effective and

functional health systems, capable of quickly and

properly answering the health issues of the

population.

1.3 Study Site

The “25 de Setembro” Healthcare Center (HC) is a

public healthcare unit attended by women and men

from the suburban areas of the Nampula Province.

Aside from the maternity and the mother and child

health wards, this HC provides internal and external

general medicine services.

The HC has only one physician, twenty five

nurses, nine midwives, eight healthcare assistants

and has 42 beds.

The maternity ward has 27 female workers.

These workers have very little technical experience,

all of the midwives have a mere elementary level

and only one of them had used a computer in the

past. None of the workers has any knowledge

regarding the use of electronic health record

systems. Despite having an Internet connection, the

signal is erratic and too weak to operate Web-based

electronic record systems.

2 OBJECTIVES

The objective of this work is to develop, implement

and evaluate a Web-based maternal and child

electronic record system in a HC located in the city

center of Nampula, Northern Mozambique, in order

to build a data repository or to improve the method

of collecting, storing, processing, accessing,

communicating and sharing information.

3 MATERIAL AND METHODS

Our approach included the following development

stages: requirements analysis; prototype

implementation; pilot testing and evaluation of the

result.

3.1 Requirements Analysis

The requirements for the prototype were defined

according to the paper-based records analysis, such

as the “antenatal forms” and the “child health card”

used in Mozambique. Accordingly, we can highlight

the following initial requirements:

 Allow access from users with low or limited

Internet speeds;

 Create forms intended for users with poor

computer skills (avoiding long steps);

 Use forms and tables in data visualization;

EVALUATION OF A MATERNAL AND CHILD ELECTRONIC HEALTH RECORD IN A DEVELOPING COUNTRY

- Preliminary Results from a Field Pilot

257

 Include in each result interface a set of queries

connected with the data;

 Build an access control, with differentiated user

permissions.

During record analysis we discovered that in

Mozambique there is no unique citizen identifier.

For that reason, we had to cross-analyze

information, using 5 fields of personal data. After

the health center record analysis, we defined the

following fields: first name of the patient; last name;

first name of the mother; first name of the father and

neighborhood of residence.

3.2 SISMI Implementation

In the Maternal and Child Health Information

System (SISMI) prototype design we used UML

language to build use case, activity, sequence and

installation diagrams, as well as the conceptual data

model.

We intentionally used Open Source technologies

in both the database and the prototype development.

The primary languages we used were PHP,

JavaScript, HTML, CSS and AJAX; we used the

MySQL database in data management.

3.3 Pilot Testing

Before the actual pilot testing, we arranged a SISMI

6 hour training program with 19 workers from the

obstetric department, in the meetings room.

In order to overcome the inevitable electricity,

Internet and management problems, we used an

offline version installed in a single computer,

equipped with a 2,10 GHZ processor, 4 Gb RAM

and a 500 Gb hard drive, and with an autonomy of 2

hours in case of power failure.

The purpose of this prototype pilot test was to

perform the initial test, to identify necessary

improvements, to adjust the system, to identify

possible problems in the EHR usage and to collect

data.

3.4 SISMI Evaluation

We evaluated the SISMI usability through

interviews and questionnaires, and the authors also

observed. The interviews were carried out in the

antenatal care room with the head of the Maternal

and Child Health program (MCH), in the maternity

ward with the maternity ward responsible and in the

registry with the director of the HC. The general

purpose of these interviews was:

 To be acquainted with the services offered and

the structure of the obstetric service in the“25 de

Setembro” HC;

 To know the opinion of the midwives, nurses and

physicians, regarding the system;

 Understand the HC culture, and socio-economic

and cultural problems.

One other activity involved the analysis of the

medical files, documentation and the observation of

the “25 de Setembro” HC, to:

 Verify existent records in antenatal records of

the pregnant woman, childbirth medical records and

existent records in the health card of the child;

 Verify the appointments log book of the pregnant

woman;

 Consult the books containing the Obstetrics

structure, the nominal list of workers, their roles,

their educational level and their allocation.

We based our questionnaire in the Likert scale, and

the workers specified their level of agreement

towards the SISMI on a 5 point scale, “1 completely

disagree”, “2 disagree”, “3 neither agree neither

disagree” 4 agree” e “5 completely agree”.

4 RESULTS

4.1 SISMI Prototype

The SISMI is a modular system composed by a

registration module, a data query module, a report

module and a statistics module.

The main menu (figure 1) grants access to

personal data of pregnant women, pregnant woman

search, pregnancy details, antenatal appointments,

childbirths, birth and appointments of the child,

displays information regarding routine check-ups of

the child and, finally, the “statistics” option displays

indicators data.

Figure 1: SISMI graphic interface.

HEALTHINF 2012 - International Conference on Health Informatics

258

4.2 Interview Results

The nurses declared that the paper-based record

maintenance in the HC was exceedingly time

consuming.

They also reported that when a pregnant woman

attends a different HC without carrying her antenatal

record, all previous information is lost or neglected

by the new physicians.

Antenatal records, and maternity records in

general, are kept in the log books of the HC, and

some of them reveal a bad conservation state. Many

of the fields of paper-based records are kept empty

or filled with assumptions made by the nurses, since

pregnant women frequently do not know the answer

or cannot remember the details, even for basic

information such as age. In some cases, nurses will

manufacture the information without even asking the

pregnant woman, erroneously assuming that she will

not know the answer. Records are rarely scrutinized

in detail afterwards, contributing to the sense of

irrelevance given to an adequate data collection.

4.3 SISMI Training

The training included a presentation of the system

and a demonstration of use. Subsequently, structured

and semi-structured exercises were performed,

simulating the reception of patients.

The initial reaction of the workers to the system

was quite positive, and they were all willing to

witness its implementation and to learn how to use

the SISMI.

The process of learning how to use the SISMI

was very difficult, and the workers seemed more

comfortable learning from their colleagues.

Although all of the workers did receive training

inside the maternity ward, only three of them were

able to effectively introduce data and use the system.

The main problems we found during training

were the misspellings, the typing speed and the

general speed when using the computer.

4.4 Pilot Testing Results

In the first day the SISMI was used several

necessary improvements were identified in the

interface, in order to enhance and simplify data

entry. The adjustment of the system to the reality of

the HC was immediately carried out by the author.

We observed that the maternity deals with a great

number of pregnant women every day, and the data

entry speed and agility of the worker were not

enough to register all women. Consequently, in

order to solve this issue, we introduced daily visits

to guide its use; additionally, we recommended that

the HC should use the off-peak hours to introduce

records from previous days, that hadn’t been

introduced yet.

It was registered 205 medical files from pregnant

women childbirths, corresponding to the month of

March, 2011. These records were entered in 11 days

of software usage, and after that stage we produced

electronic forms with some maternity indicators.

In any case, the 205 files registered in this period

represent 58% of the total 354 births that took place.

Figure 2 shows the number of appointment files

introduced in the SISMI and births per day.

Figure 2: Number of files introduced per day.

We can clearly observe a gradual increase of the

registered files, illustrating the agility improvement

in the system usage.

4.5 Record Analysis

After a record analysis we identified the introduction

of data inconsistencies in the system, such as: age,

gestational age, weight of the newly-born child, etc.

Figure 3: Age of the women at the time of delivery.

According to the data collected we concluded

that the ages registered for pregnant women ranged

from 14 to 42 years old. However, as we have

already mentioned, the ages were frequently

guessed. The age graphic for pregnant women

(figure 3), shows a small number of registered

patients with 18, 21, 26 and 31 years of age,

suggesting that these were the least ascribed ages.

Therefore, pregnant women with unknown ages

were probably aggregated in the peak categories.

Fetal maturity at the time of birth was another

insufficiently registered or blank field, and in this

1º 2º 3º 4º 5º 6º 7º 8º 9º 10º 11º

Numberof

birthsregistered

perday

Numberof

birthsperday

EVALUATION OF A MATERNAL AND CHILD ELECTRONIC HEALTH RECORD IN A DEVELOPING COUNTRY

- Preliminary Results from a Field Pilot

259

case only 190 of the 205 files displayed this

information. Table 1 shows a tendency to be born in

the 36

week (4% to 12%) and a high tendency to be

born in the 38

week (25% to 38%); accordingly,

the likelihood of being born in the 37

week could

never be 0%, as indicated. This phenomenon

suggests estimates and guesses from the staff.

Table 1: Expected delivery according to gestational age.

Gestational age (weeks) Births (N=205) [95% CI]

31 0.5% [0;2]

32 0 [0;2]

33 0 [0;2]

34 0.5% [0;2]

35 0 [0;2]

36 7% [4;12]

37 0 [0;2]

38 32% [25;38]

39 58% [51;65]

40 0 [0;2]

41 1% [0;3]

42 1% [0;3]

None of the workers remembered the last time

the scales used in birth weightings were calibrated.

Most of the midwives did not trust the values

obtained with the scales and some of them declared

they were prepared to “guess” the weight of the

infant based on their experience; this is noticeable in

the rounding of all birthweights to the hundreds and

thousands.

Figure 4 shows that more than one quarter of the

newly-born children (28%) were registered as

having exactly 3000 grams (15%) and 2500 grams

(13%); again, this suggests estimates and gross

rounding.

Figure 4: Birthweight distribution.

4.6 Usability Evaluation Results

During observation we concluded that nurses spent

less time (5 minutes) filling paper-based records,

while taking approximately 12 minutes to register

data in the SISMI.

Table 2: Usability evaluation questionnaire and median

classifications (N=9).

Questions concerning usage difficulties

N Mdn

I prefer to work with paper records 5 4

My paper record is more complete

4 2

The paper record is more convenient 4 2

I have trouble using the EHR 4 4

Questions concerning the perceived utility

EHR can be more effective in child care

5 4

EHR can improve my medical practice 8 4

EHR graphics are easy to understand

5 4

EHR is safer than the one in paper-based 5 4

EHR displayed are organized in a clear

manner

6 4

(Mdn: Median of answers)

The questionnaires, containing information

concerning the acceptability and usability of the

system, were handed to 19 workers, as well as a

detailed explanation on how to answer the

questionnaire. Only 9 of them answered the

questionnaire, and out of these only 4 answered all

questions. After the questionnaires were returned we

carried out interviews in order to understand why

some of the questions were left unanswered; several

answers followed these lines: “I was afraid that the

questionnaire would jeopardize my place in the

institution”, “I did not understand how to answer the

questionnaire” and “I did not perceive the utility of

the questionnaire”.

The analysis of the questionnaire and the

interaction with the system proved positive, and this

reveals that the majority of the staff agrees that using

this system can be beneficial to health. However,

considering the low frequency of answers and the

difficulties experienced in their filling, no real

conclusion can be derived from these answers. The

questions we presented proved to be unsuitable

given the educational level of the workers; the

evaluation method was thus inconclusive.

5 DISCUSSION

This system can guarantee the access to the first

maternal and child health indicators based in

pregnant women and serve as an example in

overcoming difficult computer conditions, like the

ones experienced in Nampula, Mozambique.

This system gave way to a simple and effective

method of creating several database query

visualizations while the access to patient data

became quicker once located in a set of database

tables. Unfortunately, the records from the HC of

Nampula are not cross-referenced, and establishing

HEALTHINF 2012 - International Conference on Health Informatics

260

associations between them is practically impossible,

leading to data redundancy and an extremely

difficult data recovery process.

The single tests we ran to check the degree of

correspondence between personal health records and

antenatal care records revealed sufficient

discrepancies to raise validity questions to any

analysis based in these records. This entails serious

implications for maternal and pediatric policy and

planning in this Province.

Despite our method, developed in order to create

a unique identifier, establishing associations across

records is still difficult, and this happens because

many women use different names according to the

purpose of their visit.

The result analysis revealed that the participants

are interested in the implementation of the SISMI.

The interaction with the SISMI proved to be positive

but inconclusive, considering the general lack of

experience with computers.

The workers identified the following benefits

connected with the SISMI and their work in the

health unit: it saves time in the monthly collection of

data; it gives way to a more detailed data analysis; it

ensures an accurate and error-free reporting process,

based in accurate; it stimulates the staff to collect

complete forms.

6 CONCLUSIONS

This system underlines the potential of health

information management in Northern Mozambique,

which can benefit millions of women and children in

the upcoming years by absorbing minimum

resources, saving both time and money.

Furthermore, this system holds unprecedented and

brand new possibilities in the healthcare quality

improvement, especially regarding maternal and

child survival.

However in Mozambique, several socio-

economic and cultural issues need to be answered

before action plans for the implementation of EHR

systems can be enacted, such as: (a) the scarcity of

human resources, both in quantity and quality; (b)

the educational level of the health sector workers,

mostly basic and intermediate; (c) inadequate

facilities; (d) high level of illiteracy; (e) insufficient

communication and transportation networks; (f) lack

of a single patient identifier.

6.1 Future Works

At present, The Nampula Provincial Health

Direction (PHD) is evaluating the proposed SISMI

with the intention of working together with us,

designing and implementing a new test in the “25 de

Setembro” HC or carrying out new pilot tests in

other HC of the Province, in order to develop a

manageable and financially sustainable program. We

also underline the development of additional

functionalities, namely: a partograph, information

regarding the prevention of vertical transmission,

implementing a unique patient identifier technology,

and so forth.

ACKNOWLEDGEMENTS

The authors would like to thank the Nampula PHD,

as well as the entire team of the “25 de Setembro”

HC. So would like to thank to the Department of

Health Information and Decision Sciences (CIDES)

in the Faculty of Medicine, University of Porto,

especially grateful to Professor Cristina Santos.

REFERENCES

(2005). Maternal Mortality in 2005: Estimates developed

by WHO, UNICEF, UNFPA, and The World Bank. In.

(2007). Moçambique - Ministério da Saúde: Relatório

sobre as Boas Práticas do Sistema de Informação em

Saúde de Moçambique. In.

(2008). Sistemas de Informação Sanitária em África:

Recursos, Indicadores, Tratamento de Dados,

Divulgação e uso. In: Conferência Ministerial de

Argel sobre Investigação em Saúde na Região

Africana.

(2009a). Moçambique - Ministério da Saúde, OMS,

UNFPA, UNICEF, USAID: Parceria Nacional para a

Promoção da Saúde Materna, Neonatal e Infantil. In.

(2009b). Moçambique - Ministério da Saúde: Plano

Estratégico do Sistema de Informação para a Saúde

(SIS) - 2009-2014. In, Direcção de Planificação e

Cooperação Maputo - Moçambique.

Allotey, P. A., and Reidpath, D., (2000). Information

quality in a remote rural maternity unit in Ghana. In:

Health Policy Plan 15, 170-176.

Berkowicz, D. A., Chueh, H. C., and Barnett, G. O.,

(1997). Design considerations in migrating an

obstetrics clinical record to the Web. In: Proc AMIA

Annu Fall Symp, 754-758.

Dumont, A., Tourigny, C., and Fournier, P., (2009).

Improving obstetric care in low-resource settings:

implementation of facility-based maternal death

reviews in five pilot hospitals in Senegal. In: Hum

Resour Health 7, 61.

Elizabeth, M., Cristiano, M., Ana M., L., Harriet, B., and

Ezekiel, K., (2007). Mulher, sida e o acesso à saúde na

EVALUATION OF A MATERNAL AND CHILD ELECTRONIC HEALTH RECORD IN A DEVELOPING COUNTRY

- Preliminary Results from a Field Pilot

261

áfrica subsahariana, sob a perspectiva das ciências

sociais. In: vita.

Fraser, H. S., Biondich, P., Moodley, D., Choi, S.,

Mamlin, B. W., and Szolovits, P., (2005).

Implementing electronic medical record systems in

developing countries. In: Inform Prim Care 13, 83-95.

Gladwin, J., Dixon, R. A., and Wilson, T. D., (2002).

Rejection of an innovation: health information

management training materials in east Africa. In:

Health Policy Plan 17, 354-361.

Hayrinen, K., Saranto, K., and Nykanen, P., (2008).

Definition, structure, content, use and impacts of

electronic health records: a review of the research

literature. In: Int J Med Inform 77, 291-304.

Kalogriopoulos, N. A., Baran, J., Nimunkar, A. J., and

Webster, J. G., (2009). Electronic medical record

systems for developing countries: review. In: Conf

Proc IEEE Eng Med Biol Soc 2009, 1730-1733.

NEPAD, (2008). Estratégia para a Saúde. In: A Nova

Parceria para o Desenvolvimento de África.

OMS, (2010). Análise dos Recursos Humanos da Saúde

nos Paises Africanos de Lingua Oficial Portuguesa

[PALOP]. In.

Slagle, T. A., (1999). Perinatal information systems for

quality improvement: visions for today. In: Pediatrics

103, 266-277.

Zandieh, S. O., Yoon-Flannery, K., Kuperman, G.J.,

Langsam, D. J., Hyman, D., and Kaushal, R., (2008).

Challenges to EHR implementation in electronic-

versus paper-based office practices. In: J Gen Intern

Med 23, 755-761.

HEALTHINF 2012 - International Conference on Health Informatics

262