HMIS DATA QUALITY

DATA QUALITY
a major concern for large companies especially in the areas of customer relationship management (CRM), data integration, and regulation requirements.
Poor data quality generates cost, affects customer satisfaction, company reputation, and even the strategic decisions of the management.
The overall utility of a dataset(s) as a function of its ability to be processed easily and analyzed for a database, data warehouse, or data analytics system

According to Rouse (2005), data quality signifies the data’s appropriateness to serve its purpose in a given context. Having quality data means that the data is useful and consistent. Data cleansing can be done to raise the quality of available data.

Lot Quality Assurance Sampling (LQAS)

a tool that allows the use of small random samples to distinguish between different groups of data elements (or lots) with high and low data quality.
For health managers and supervisors, using small samples makes the conduct of surveys more efficient.
This tool had been widely applied in the healthcare industry for decades and has been primarily used for quality assurance of products
The concept and application of LQAS have been adopted in the context of District Health Information System (DHIS) data quality assurance. The adoption was comprised of designating health facilities, monthly reports, sections of monthly reports, and group of data elements as ‘lots’ to provide representative samples for data quality assurance of DHIS

STEPS IN APPLYING LQAS
Define the service to be assessed (e.g. DQA of DHIS)
Identify the unit of interest (e.g. a supervisory area, facility, hospital, a district).
Define the higher and lower thresholds of performance based on prior information about the expected performance of the region of interest.
Determine the level of acceptable error.
Determine the sample size and decision rule for acceptable errors to declare an area as performing “below expectations”.
Identify the number of errors observed (mismatched data elements will be reliably determined if the facility is performing above or below expectations).

ROUTINE DATA QUALITY ASSESSMENT (RDQA)

A simplified version of the Data Quality Audit (DQA) tool which allows programs and projects to verify and assess the quality of their reported data.
Aims to strengthen data management and reporting systems.
Objectives of RDQA:
Rapidly verify the quality of reported data for key indicators at selected sites.
Implement corrective measures with action plans for strengthening data management reporting systems and improving data quality.
Monitor capacity improvements and performance of data management and reporting system to produce quality data.
The RDQA is a multipurpose tool that is most effective when routinely used.

Uses for the RDQA tool:
Routine data quality checks as part of on-going supervision
Routine data quality checks can be included in already planned supervision visits at the service delivery sites.
Initial and follow-up assessments of data management and reporting systems
Repeated assessments (e.g. biannually or annually) of a system’s ability to collect and report quality data at all levels can be used to identify gaps and monitor necessary improvements
Strengthening of the program staff’s capacity in data management
Monitoring and evaluation (M&E) staff can be trained on the RDQA and sensitized to the need to straighten the key functional areas linked to data management and reporting in order to produce quality data.
Preparation for a formal data quality audit
The RDQA tool can help identify data quality issues and areas of weakness in the data management and reporting system that would need to be strengthened to increase readiness for a formal data quality audit
External assessment by partners of the quality of data
Such use of the RDQA for external assessments could be more frequent, more streamlined, and less resource intensive than comprehensive data quality audits that use the DQA version for auditing.

DEVELOPMENT IMPLEMENTATION PLAN
An implementation plan is a project management tool that illustrates how a project is expected to progress at a high level. It helps ensure that a development team is working to deliver and complete tasks on time (Visual Paradigm, 2009).
Important in ensuring the efficient flow of communication between those who are involved in the project as to minimize issues that would delay delivery of the project.
Validates the estimation and schedule of the project plan.
An implementation plan is developed through the following key steps (Smartsheet, 2017)
Define goals/objectives
Address the question, “What do you want to accomplish?”
Schedule milestones
Outline the deadline and timelines in the implementation phase.
Allocate resources
Determine whether you have sufficient resources, and decide how you will procure those missing.
Designate team member responsibilities
Create a general team plan with overall roles that each team member will play
Define metrics for success
How will you determine if you have achieved your goals?

DATA QUALITY TOOLS
Analyzes information and identifies incomplete or incorrect data.
Data cleansing follows after the complete profiling of data concerns, which could range anywhere from removing abnormalities to merging repeated information.
By maintaining data integrity, the process enhances the reliability of the information being used by an organization. Usually, these data quality software products can share features with master data management, data integration, or big data solutions.
Gartner (2017) explains how these data quality tools are used to address problems in data quality:
Parsing and standardization
– the decomposition of fields into component parts and formatting the values into consistent layouts based on industry standards and patterns and user-defined business rules .
Generalized “cleansing”
– the modification of data values to meet domain restrictions, constraints on integrity, or other rules that define data quality as sufficient for the organization.
Matching
– the identification and merging of related entries within or across data sets
Profiling
– the analysis of data to capture statistics or metadata to determine the quality of the data and identify data quality issues.
Monitoring
– the development of controls to ensure conformity of data to business rules set by the organization.
Enrichment – the enhancement of the value of the data by using related attributes from external sources such as consumer demographic attributes or geographic descriptors.
As data quality continues to become increasingly all-encompassing, data integration tools are further developed to include data quality management functionality.

APPLICATION/SCOPE OF DATA QUALITY TOOLS
The first generation of data quality tools was characterized by dedicated data cleansing tools designed to address normalization and reduplication. However, in the last 10 years, it was observed that there is a generalization of Extract, Transform, Load (ETL) tools which allow the optimization of the alimentation process.
Recently, these tools started to focus on Data Quality Management (DQM), which generally integrates profiling, parsing, standardization, cleansing, and matching processes (Goasduez Nugier, Duquennoy, and Laboisse, 2007)

ROOT CAUSE ANALYSIS
A problem solving method that identifies the root causes of problems or events instead of simply addressing the obvious symptoms.
The aim is to improve the quality of products and services by using systematic ways to address problems in order to be effective (Bowen, 2011)

TECHNIQUES IN ROOT CAUSE ANALYSIS
Root cause analysis is among the core building blocks in the continuous improvement efforts of an organization in terms of its operation dynamics, especially in the way it handles information.
However, root cause analysis alone will not produce any valuable results.
The organization should seek to improve at every level and in every department for this to work.
The analysis will help develop protocols and strategies to address underlying issues and reduce future errors.
Bowen (2011) suggests that “to address the root cause of a problem, one must identify the problem and ask “why” five times to determine the proper strategies to address its root cause”.

Failure Mode and Effects Analysis (FMEA)
Aims to find various modes of failures within a system and addresses the following questions for execution:
What is the mode in which an observed failure occurs?
How many times does a cause of failure occur?
What actions are implemented to prevent this cause from occurring again?
Are these actions effective and efficient?
FMEA is used when there is a new product or process it when there are changes or updates in a product and when a problem is reported through customer feedback.

2. Pareto Analysis
Uses the Pareto principle which states that 20 percent of the work creates 80 percent of the results.
Used when there are multiple causes to a problem.
Pareto chart in an Excel software: lays down the potential causes in a bar graph and tracks the collective percentage in a line graph to the top of the table.
The reflected causes from the table should account for at least 80 percent of those involved in the analysis.

3. Fault Tree Analysis (FTA)
Used in risk and safety analysis.
Uses boolean logic to determine the root causes of an undesirable event.
The undesirable result is listed at the top of the tree and then all the potential causes are listed down to form the shape of an upside down tree.

4. Current Reality Tree (CRT)
Used when the root causes of multiple problems need to be analyzed all at once.
The problems are listed down followed by the potential cause for a problem.
By doing so, a cause common to all problems will appear.

5. Fishbone Diagram
Also call the Ishikawa or cause-and-effect
Looks like a fishbone to show the category Small and clear
Useful in grouping causes (e.g. people, measurements, method, materials, environment (machine) into category.
Categories should be 4Ms (manufacturing); 4S(service): 8 Fs (also service) depending on the industry.

6. Kepner-Tregoe
Breaks a problem down to its roots cause by assessing a situation using priorities and orders or concern for specific issues.
The various decisions that should be made to address the problem and then outlined.
Then a potential analysis is made to ensure that the actions recommended are sustainable.

7. Rapid Problem Resolution (RPR Problem Diagnosis)
Diagnoses the causes of recurrent problems by following the three phases below:
Discover – data gathering and analysis of findings
Investigate – creation of a diagnostic plan and identification of the root cause through careful analysis of the diagnostic data
Fix – fixing the problem and monitoring to confirm and validate that the correct root cause was identified.

SUSTAINING A CULTURE OF INFORMATION USE

Information culture affects information use. Information culture is determined by the following variables: mission, history, leadership, employee traits, industry, and national culture.
Can be shaped by cognitive and epistemic expectations which are influenced by the way tasks are performed and decisions are made
The result suggests that in order to have a sense of information attitude and values, managers should consider taking the pulse of information of their own organization. The sets of identified behaviors and values could account for significant proportions of the variance in information use.
Management plays an important role is maintaining and improving the quality of data and information used in daily operations.

KEY POINTS TO REMEMBER
Data quality is the overall utility of a dataset as a function of its ability to be processed easily and analyzed for a database, data warehouse, or data analytics system.
LQAS is a tool that allows the use of small random samples to distinguish between different groups of data elements (or lots) with high and low data quality.
RDQA tool is a simplified version of the DQA which allows programs and projects to verify and assess the quality of their reported data.
The development of an implementation plan is important in ensuring that the communication between those who are involved in the project will not encounter any issues and work will also be delivered on time.
A root cause analysis is a problem solving method aimed at identifying the root causes of problems or events instead of simply addressing the obvious symptoms.
Techniques in root cause analysis include failure mode and effects analysis (FMEA), Pareto Analysis, FTA, CRT, Fishbone or Ishikawa, Kepner-Tregoe technique, and RPR problem diagnosis.