p–ISSN: 2723 – 6609 e-ISSN: 2745-5254
Vol. 5, No. 12, December 2024 http://jist.publikasiindonesia.id/
Indonesian Journal of Social Technology, Vol. 5, No. 12, December 2024 6060
Case Study of Claim Data and Participant Data in Indonesian
Insurance Companies
Chris Solontio1, Achmad Nizar Hidayanto2
Universitas Indonesia, Indonesia
Email: [email protected]*, [email protected]
*Correspondence
ABSTRACT
Keywords: data
management; data quality;
health insurance; data
quality dimensions.
Data has become an asset for insurance companies,
especially health insurance, which has many benefits so
management needs to realize the importance of good data
quality to avoid the impact of poor data quality. In this study,
data quality measurement will be carried out by observation
to see the total amount of invalid data from the data
dimensions, namely, accuracy, completeness, and
consistency on the connection between claim data and
participant data as well as findings from all data from the
case study site of this research. In addition to the data
analysis, interviews were conducted with the IT, Customer
Retention, Operations, and Actuarial teams which are
directly related to data flow and data processing to make
analyses that assist management in making decisions. From
the results of the analysis and interviews that have been
conducted, there are still inappropriate data and obstacles
faced by users in dealing with poor data quality. From the
results of the analysis, management needs to form a data
govenance team that has responsibility for the entire data
flow and maintains data quality. Later, the data set that has
been managed will have a positive impact on other teams in
terms of analyzing trends or fraud in a faster time, helping in
the creation of data warehouses, the application of artificial
intelligence (AI) and digital transformation as a form of
company improvement to insurance policyholders.
Introduction
According to the results of the Populix survey, the majority of Indonesians already
have BPJS Kesehatan insurance of 83% and private insurance 38% (Data Indonesia, n.d.).
One of the most owned private insurances is health insurance (Katadata, N.D.). This
makes the medical data record on the use of health insurance will increase over time. The
increase in health data in insurance companies makes management want to use the data
set to provide improved services to participants, carry out digital transformation or
competitive business competition such as product offerings by looking at the
segmentation of participants or providing the best premiums. Therefore, one of the things
Chris Solontio, Achmad Nizar Hidayanto
Indonesian Journal of Social Technology, Vol. 5, No. 12, December 2024 6061
that every company must have is the quality of the data that is recapped every day, has
good data quality, no data silos and is well integrated between claim data, products, and
participant data. From what has been explained, this study will use one of the life
insurance companies to assess the quality of the data owned by the life insurance
company.
This life insurance company is one of the largest life insurance companies in
Indonesia that has a wide range of hospital or clinic partners and has two main health
products, namely managed care and indemnity. Both health products provide health
protection for employees who work in a company. In this study, the quality of the data
from the managed care product will be assessed, where this product is a tiered health
insurance product whose services must go through a first-level health facility, namely a
clinic or health center, then if further action is needed, a referral will be given to the
hospital for outpatient treatment at the advanced level or inpatient with a note that the
diagnosis and services provided to the participants need to be recorded. From each of
these steps, each provider needs to create an SJP number to be used as a unique number
from the medical recap journey from the first level, advanced level to the need for
hospitalization carried out by the participant.
In the face of today's business challenges, the company has the initiative to improve
its services to be the best and help management quickly make decisions through digital
transformation. Some of these initiatives have been carried out such as accelerating the
claim process, creating dashboards for the company's internal and assisting management
in making decisions from cash flow results, claim behavior and others. Meanwhile,
initiatives that have not been implemented are the creation of a data warehouse, and the
use of artificial intelligence (AI) in detecting fraud. From the initiative that has been
running, there are obstacles where the process takes approximately 3 hours because it
needs to validate data because there is still incomplete, siloed and inconsistent data.
Meanwhile, initiatives that have not been implemented are also experiencing obstacles
because they cannot be executed due to poor data conditions. Therefore, there are
obstacles in data quality.
Data quality is one of the components in data management that can produce
efficient operational processes, help decision-making, create data warehouses, and have
a positive influence on customer satisfaction (David, 2011). In maintaining data quality,
every company needs to define data dimensions to know the impact of poor data quality
on costs, reputation, compliance regulations, and so on (Askham, Cook, Doyle, Fereday,
& Gibson, 2013). In the insurance business line, poor data quality in general can lead to
losses in operational and strategic costs for hidden costs and direct costs (Haug,
Zachariassen, and van Liempd, 2011). hidden costs and direct costs from the effects of
Case Study of Claim Data and Participant Data in Indonesian Insurance Companies
Indonesian Journal of Social Technology, Vol. 5, No. 12, December 2024 6062
poor-quality data on operational activities and strategies for insurance companies as
shown in Figure 1.
Figure 1
Consequences of Poor Data Quality in Insurance
In data quality for the information age, Thomas Rednan formulated a set of data
quality dimensions that are rooted in data structures. In addition, the dimensions of data
cannot be determined the same in every business area but can vary in each company
depending on the characteristics of the data or the use of data by the company. The
importance of data dimensions as a tool to measure data quality in an organization, many
researchers conduct this research by conducting literature reviews or case studies in
several business lines such as Table 1.
Chris Solontio, Achmad Nizar Hidayanto
Indonesian Journal of Social Technology, Vol. 5, No. 12, December 2024 6063
Table 1
Data Quality Research Area
Data Quality
Research Area
Reference
Government Government
Organization (Nulhusna,
Taufiq, & Ruldeviyani,
2022), BPS-Statistics
Indonesia (Rahmawati
and Ruldeviyani, 2019),
Malaysian Public
(Zarina Abdul Jabar et
al., 2022), State
Electricity Company
(Purnomoadi et al.,
2023).
Financial
industry
PT BPI (Sunandar and
Nizar Hidayanto, 2022).
Airport services PT JAS (Wahyudi and
Isa, 2023).
Education Institute of Statistics
(Wijayanti et al., 2018).
Factory Paper Factory
(Xu, Zhang, and Shi,
2022).
Medical or
health
Electronic Medical
Record (Miguel et al.,
2021).
Method
In assessing the quality of this data, observations will be made with profiling that
has been carried out in Mary's research by forming several roles that are carried out and
looking at the number of invalid data that occurs in claim data and participant data by
paying attention to the dimensions of data quality that will be used based on data
characteristics, namely accuracy, completeness, and consistency. In addition, the
interview process was carried out in several related units whose work activities are
directly related to claim data and participants to find out the causes of poor data quality
and the resulting impact.
The limitation of this study is that the claim data analyzed is claim data that
occurred at a provider (hospital/clinic) in 2022 and the amount of data provided was 125
thousand data consisting of 38 thousand SJP data, 9 thousand claim participants, and three
providers. SJP is a letter of guarantee for patients after health services are carried out at a
hospital or clinic. In data processing, it will use excel and Python to view invalid data
from participant data and claim data.
Case Study of Claim Data and Participant Data in Indonesian Insurance Companies
Indonesian Journal of Social Technology, Vol. 5, No. 12, December 2024 6064
Results and Discussion
Based on the six roles that have been formed from claim data and participant data
to assess the quality of the data, invalid data from each data dimension is found as shown
in Table 3. The results were:
1. For the first role, no invalid data was found on the receipt date and delivery date data
by looking at the length of stay.
2. Participant Status with a diagnosis was found to be 2 SJPs invalid with information on
the status of child participants but the diagnosis was Z34 which is a diagnosis of normal
pregnancy surveillance.
3. Based on the list of diagnoses per gender based on ICD X on the ICD10CM website,
it was found that 13 SJPs were invalid in the data claiming a relationship between
gender and diagnosis (Table 4).
4. The type of service with a diagnosis of 12,307 SJP is inconsistent where the service
does not have details of the type of medication administered.
5. There is an inconsistency between the participant plan and the participant data where
there are 36 participants who have 2 different plans for example participant number
00001 but have silver and blue insurance plans
6. It was found that 14 participants whose age range did not match their membership
status (Table 2).
Table 2
Invalid Data on ICD X Diagnosis
By Gender
Gender Diagnose Invalid
Male O74, N91, N89, N80, and N93
Female C61, D17.6, and N47
Table 3
Invalid data on age range
Based on Participant Status
Status Age Range Total
Participants
spouse
0 - 5 1
6 - 12 2
13 - 16 6
child
36 - 45 4
46 - 55 1
In addition to the results of data analysis, interviews were conducted with
departments that work directly or manage data, namely information technology, customer
retention, operations, and economics to find out the effects and consequences of poor data
quality. From the results of the interview, several points were found as follows:
Chris Solontio, Achmad Nizar Hidayanto
Indonesian Journal of Social Technology, Vol. 5, No. 12, December 2024 6065
Table 4
Total Invalid Data Based on Data Quality Dimension
It Roles Accuracy
Complete
ness
Consistency
1. Service date and
discharge date
(hospitalization)
0 0 0
2. Participant
status with ICD
X Diagnosis
2 0 0
3. Gender with
ICD X
Diagnosis
13 0 0
4. Types of
services with
ICD X
Diagnosis
0 0 12.307
5. Participant plan
and participant
number
0 0 36
6. Participant
status and age
range
14 0 0
Information Technology
There is no single department in the IT field that is specifically responsible for
maintaining the quality of the data in the system. Although for now the IT service team
and business analyst work together in carrying out their duties as MIS and data engineers,
this makes the team overwhelmed because there is a double task of the main task of the
two teams in maintaining the existing data, sometimes only relying on findings from users
if anomalous data is found when processing data and in terms of validation it also takes a
long time.
Customer Retention
Anomalies in the date of birth data of participants and their family members caused
by the absence of format locking on the date of birth from a file manually uploaded to the
system caused an error in the participant's date of birth.
Operational
The data processing for the analysis material takes a long time because inconsistent
and non-uniform data are found, especially in the outpatient or inpatient service detail
codes caused by each hospital with different inputs does not have a uniform code for the
service detail code.
Actuarial
It takes a long time in the data validation section because there is an inconsistency
between participant data and claim data, so it must be validated by the data owner or IT
who provides the data. This has an impact on determining premium extensions for
participants who are not fast and analyzing claims for additional purposes of audit reports,
as well as providing recommendations and information on inflation that occurs to the
pricing team in designing a baseline percentage of one of the components of premium
calculation and other reports needed by the actuarial team.
Case Study of Claim Data and Participant Data in Indonesian Insurance Companies
Indonesian Journal of Social Technology, Vol. 5, No. 12, December 2024 6066
From the results of the observations and interviews, this data quality problem is
caused by:
1. no team is fully responsible for the data. Although at this time the service IT team is
appointed to be responsible for the data on the server, it is not optimal because of the
double work carried out by the service IT team and business analysts, it adds to the
tasks of the related team as the person in charge of data as a task as an MIS team or
data engineer.
2. There are no rules for validation in the system. One of them is the existence of gender-
inappropriate age data and inconsistencies in a criterion, as stated by the operational
team that there is no format locking for the date of birth of the Excel file that is
manually uploaded to the system, causing errors in the participant's date of birth.
3. There is no standardization of service detail code data from the data that comes in from
the hospital to the company's system that has been provided. This happens because the
data has not been standardized by the operational team to IT. This needs to be done
immediately because uniform service details will help easier analysis for data working
units such as underwriting and pricing teams and operations teams.
From point 1, there are no regulations in the company that govern the overall data
management to avoid poor data quality.
Conclusion
After analyzing the data quality, it was found that the data quality problem arose
due to the lack of awareness from the management and the absence of a data management
team that was specifically responsible for all the data elements or outputs that would be
generated from the company's data. This needs to be a special concern for management
because insurance companies are financial companies like banks where data has become
important to improve services and avoid fraud. From these problems, there are three
suggestions to help insurance companies in improving the quality of their data, namely:
1. In answering the problems faced by the IT team, the IT team needs a special team to
pay attention to data consisting of data engineers and data stewards. According to
DMBOK, this data team has responsibility for all important elements in data
management in the company where its tasks include creating data architecture, data
quality standards, metadata, data file documentation against data definitions or
criteria, updating master data or reference data, data warehouse, and business
intelligence.
2. The next stage is if a data governance team has been formed, the company can use a
vendor such as Collibra, or create its own system that will maintain and directly
identify incoming data to maintain good data quality.
3. Because there is already a dashboard created, it is necessary to create a data
warehouse in order to speed up the process of creating a dashboard or creating a data
mart for every user who wants to use data from a direct database.
Chris Solontio, Achmad Nizar Hidayanto
Indonesian Journal of Social Technology, Vol. 5, No. 12, December 2024 6067
Bibliography
Askham, N., Cook, D., Doyle, M., Fereday, H., & Gibson, M. (2013). The Six Primary
Dimensions For Data Quality Assessment. DAMA International.
Data Indonesia. (n.d.). Survei Mayoritas Warga Indonesia Miliki Asuransi Kesehatan.
David, Loshin. (2011). The Practitioner’s Guide to Data Quality Improvement.
Knowledge Integrity, Inc.
Haug, Anders, Zachariassen, Frederik, & van Liempd, Dennis. (2011). The costs of poor
data quality. Journal of Industrial Engineering and Management, 4(2), 168–193.
https://doi.org/10.3926/jiem.2011.v4n2.p168-193
katadata. (n.d.). Ini Jenis Asuransi Swasta yang Paling Banyak Dimiliki Penduduk
Indonesia.
Miguel, Diaz Iturry, Solange N., Alves Souza, Marcia, Ito, & Da Silva Suzana Alves.
(2021). Data Quality in health records: A literature review. Iberian Conference on
Information Systems and Technologies, CISTI, (June), 1–6.
https://doi.org/10.23919/CISTI52073.2021.9476536
Nulhusna, Rizqa, Taufiq, Nur Fajar, & Ruldeviyani, Yova. (2022). Strategy to Improve
Data Quality Management: A Case Study of Master Data at Government
Organization in Indonesia. Proceeding - 2022 International Symposium on
Information Technology and Digital Innovation: Technology Innovation During
Pandemic, ISITDI 2022, 150–155.
https://doi.org/10.1109/ISITDI55734.2022.9944466
Purnomoadi, Andreas, Sari, Ira M., Anna Maria, J. S., Fiddiansyah, Dimas B., Saputro,
Nuriyanto E., & Sofan Hadi, M. S. (2023). A Method to Quantify Data Quality in
Asset Health Indices Model. Proceedings of 2023 4th International Conference on
High Voltage Engineering and Power Systems, ICHVEPS 2023, 16–20.
https://doi.org/10.1109/ICHVEPS58902.2023.10257386
Rahmawati, Sinta Denovi, & Ruldeviyani, Yova. (2019). Data Quality Management
Strategy to Improve the Quality of Worker’s Wage and Income Data: A Case Study
in BPS-Statistics Indonesia, 2018. Proceedings of 2019 4th International
Conference on Informatics and Computing, ICIC 2019.
https://doi.org/10.1109/ICIC47613.2019.8985803
Sunandar, Nandang, & Nizar Hidayanto, Achmad. (2022). Data Quality Management
Improvement: Case Studi PT BPI. 2022 IEEE World AI IoT Congress, AIIoT 2022,
53–58. https://doi.org/10.1109/AIIoT54504.2022.9817195
Wahyudi, Tedi, & Isa, Sani Muhammad. (2023). Data Quality Assessment Using Tdqm
Framework: a Case Study of Pt Aid. Journal of Theoretical and Applied
Information Technology, 101(9), 3576–3589.
Case Study of Claim Data and Participant Data in Indonesian Insurance Companies
Indonesian Journal of Social Technology, Vol. 5, No. 12, December 2024 6068
Wijayanti, Wiluyaningtyas, Hidayanto, Achmad Nizar, Wilantika, Nori, Adawati, Infaz
Rizki, & Yudhoatmojo, Satrio B. (2018). Data quality assessment on higher
education: A case study of institute of statistics. 2018 International Seminar on
Research of Information Technology and Intelligent Systems, ISRITI 2018, 231–
236. https://doi.org/10.1109/ISRITI.2018.8864476
Xu, Dongting, Zhang, Zhisheng, & Shi, Jinfei. (2022). A Data Quality Assessment and
Control Method in Multiple Products Manufacturing Process. 2022 5th
International Conference on Data Science and Information Technology, DSIT 2022
- Proceedings, 1–5. https://doi.org/10.1109/DSIT55514.2022.9943883
Zarina Abdul Jabar, Zam, Wook, Muslihah, Zakaria, Omar, Ramli, Suzaimah, & Afiza
Mat Razali, Noor. (2022). Establishment of Big Data Analytics Application Model
for Malaysian Public Sector: an Expert Validation. Edpacs, 65(4), 1–20.
https://doi.org/10.1080/07366981.2021.1958736