18.2 Health Information Systems, Records & KHIS Reporting
Key Takeaways
- Kenya's national health information system is the Kenya Health Information System, built on DHIS2, into which facility summary data flows through the sub-county and county levels.
- Primary data is captured in MOH registers at the point of care and aggregated into monthly summary tools, including MOH 705A and 705B for outpatient morbidity, MOH 710 for immunisation, MOH 711 as the integrated service summary, and MOH 717 for hospital workload.
- The routine reporting rhythm is that facilities submit by around the 5th of the following month and sub-counties enter data into KHIS by the 15th, which is the date against which national reporting timeliness is scored.
- Data quality is assessed on completeness, timeliness, accuracy, consistency, and integrity, and a data quality audit works by recounting from the source register and comparing that count with what was reported.
- The Digital Health Act 2023 provides the legal framework for health information systems, record interoperability, e-health, and the protection and retention of health data.
18.2 Health Information Systems, Records & KHIS Reporting
Quick Summary: Every figure on a national health dashboard began as a line a clinical officer wrote in a register. The chain is register → monthly summary tool → sub-county → KHIS (DHIS2) → county and national dashboards, on a rhythm of facility submission by about the 5th and KHIS entry by the 15th. Learn the tool numbers, the deadlines, and the five dimensions of data quality.
Why the Register Entry Matters
Routine health information in Kenya is used for four things: planning (how many staff and commodities a facility needs), resource allocation (county budget lines follow reported burden), monitoring (whether coverage targets are being met), and outbreak detection (an unusual rise in a reported condition). All four depend on an entry made in a consulting room by someone under time pressure.
The governing principle: data quality is decided at the point of capture. Nothing downstream — no audit, no dashboard, no analysis — can recover information that was never accurately recorded. A facility that under-reports its malaria burden will be allocated fewer antimalarials, and the people harmed by that will be its own patients.
The Two Layers: Registers and Summary Tools
Primary (source) tools are completed at the moment of care — patient cards and registers. Secondary (summary) tools aggregate those entries at the end of the month for reporting.
| Tool | What it is |
|---|---|
| MOH 204A / 204B | Outpatient register (under-5 and over-5) — primary capture of every outpatient encounter |
| MOH 705A | Monthly outpatient morbidity summary, under 5 years |
| MOH 705B | Monthly outpatient morbidity summary, 5 years and over |
| MOH 710 | Immunisation and vitamin A summary |
| MOH 711 / 711A | Integrated summary tool for service delivery, including reproductive health |
| MOH 717 | Monthly workload report (hospital) |
| MOH 240 | Laboratory register |
| MOH 731 | Comprehensive HIV and ART services report |
| MOH 505 | Weekly notifiable disease and IDSR reporting |
| MOH 363 | Post-rape care / sexual violence management form |
Learn the 705A / 705B pair. Splitting outpatient morbidity by under-5 and 5-and-over is the most frequently examined tool distinction, and these two forms also carry the highest reporting rates in the national system, so they are the ones a supervisor will open first.
The Reporting Chain and Its Deadlines
POINT OF CARE REGISTERS (MOH 204A/B, 240, 512 …)
│ daily entry at the moment of care
▼
FACILITY SUMMARY TOOLS (705A/B, 710, 711, 717)
│ compiled at month end and submitted ── by about the 5th of the next month
▼
SUB-COUNTY HEALTH RECORDS OFFICER Verifies, compiles, enters into KHIS
│ ── by the 15th of the next month
▼
COUNTY Review, feedback, county dashboards
│
▼
NATIONAL — KHIS (DHIS2) Indicators, dashboards, national planning
The 15th is the date that counts. National reporting timeliness is scored against it, which means a facility that submits on the 12th has not made the deadline late — it has made it just in time for someone else to do their part.
Feedback is the weakest link in the chain and the most examinable. Data that flows only upward demoralises the people collecting it and degrades quality over time. A functioning system returns analysis downward — county dashboards, comparative performance, and discussion of the facility's own numbers at supervision. If a vignette describes staff who "fill the forms but nobody ever tells us anything," the system deficiency is absent feedback, and the fix is a scheduled data review meeting at facility level.
The Five Dimensions of Data Quality
| Dimension | Question | Typical failure |
|---|---|---|
| Completeness | Were all expected reports and all fields submitted? | A facility reports nine months out of twelve; a tool submitted with blank sections |
| Timeliness | Did the report arrive by the deadline? | Report submitted on the 20th, after the KHIS cut-off |
| Accuracy | Do reported figures match the source register? | Summary shows 340 outpatient cases; recount of the register gives 290 |
| Consistency | Do related figures agree with each other and over time? | More children reported as fully immunised than the total reported as attending |
| Integrity | Is the data free from deliberate manipulation? | Figures inflated to hit a coverage target or to trigger a commodity allocation |
How a Data Quality Audit Works
A data quality audit does exactly one thing conceptually: it recounts from the source and compares.
- Select an indicator and a reporting period.
- Recount the cases from the primary register for that period.
- Compare that recount with the figure reported in the summary tool and with the figure in KHIS.
- Compute the verification factor: recounted value ÷ reported value. A factor of 1.0 is perfect agreement; below 1.0 indicates over-reporting and above 1.0 indicates under-reporting.
- Investigate the cause, agree corrective action, and retrain where needed.
Note the direction of the logic carefully, because it is a favourite trap: a recount that is lower than the reported figure means the facility over-reported.
Constructing an Indicator Correctly
Most reported health indicators are proportions, and most errors are denominator errors.
For coverage indicators the denominator is normally a catchment population estimate — for example, surviving infants for immunisation coverage, or expected pregnancies for antenatal coverage — and not the number of people who happened to attend the facility. A facility that uses attendees as the denominator will report coverage close to 100% while missing every child who never came, which is precisely the group the indicator exists to find.
The National Platforms
| Platform | What it holds |
|---|---|
| KHIS (Kenya Health Information System, DHIS2) | Routine aggregate service delivery data — the national reporting backbone |
| KMHFL (Kenya Master Health Facility List) | The authoritative register of health facilities, with the facility code used in reporting and in SHA empanelment |
| eCHIS (Electronic Community Health Information System) | Community-level data captured by Community Health Promoters at household level |
| IHRIS | Human resources for health information |
| DamuKE | Blood transfusion and blood bank management |
| KHRO (Kenya Health and Research Observatory) | Analytical and research data products |
The Legal Duties Attaching to Records
The Digital Health Act 2023 supplies the legal framework for health information systems, interoperability of records between facilities, recognition of e-health as a mode of service delivery, and the protection and retention of health data — the Act contemplates retention of an individual's health data for a substantial minimum period, with disposal governed by regulations. Alongside it, the Data Protection Act 2019 classifies health data as sensitive personal data requiring a lawful basis for processing.
At facility level the practical duties are unglamorous and examinable:
- The physical record belongs to the facility; the information belongs to the patient. A patient may access their own information, but does not simply remove the file.
- Confidentiality applies to the record as much as to conversation. Files left open on a desk, results discussed in a corridor, and unlocked registers are all breaches.
- Entries are contemporaneous, legible, dated, timed, and signed, with the writer identifiable.
- Never erase or obliterate. Correct an error by striking a single line through it so that the original remains readable, then write the correction, date, and initial it. An obliterated entry looks like concealment and destroys the evidential value of the whole record.
- Retain records for the statutory period and store them securely; do not dispose of them informally.
A sub-county health records officer is compiling monthly returns. Which tool provides the monthly summary of outpatient morbidity for children under five years of age?
A data quality audit recounts antenatal first visits from the facility ANC register for a given month and finds 290 visits. The facility reported 340 for the same month. What does this indicate, and what is the verification factor?
Staff at a dispensary complain that they complete registers and summary tools every month but have never once been shown any analysis of their own data or told how their facility compares with others. Which health information system function is failing?
A clinical officer realises that a drug dose recorded in a patient file an hour earlier was written incorrectly. What is the correct way to amend the record?