# Codebook — Rural Health Statistics of India (harmonised, 2005–2023) A datasheet for the harmonised RHS dataset: what it contains, how it was built, what the values mean, and how to use it responsibly. This documentation is open even while the dataset files are shared on request (see *Access*). **Version 1.0 · 2026 · Association for Socially Applicable Research (ASAR)** --- ## 1. Overview | | | |---|---| | Indicators | 863 canonical indicators | | Domains | 12 (Demographics, Facilities, Workforce Rural, Workforce Urban, Infrastructure, Tribal, Coverage, Training, Five-Year Plan, FRU, HWC, Other) | | Geography | 36 current States & Union Territories (40 including historical entities) | | Time | 19 annual rounds, 2005 to 2022-23 | | Observations | 2,31,597 (long format: one row per Year × State × Indicator) | | Source | Rural Health Statistics annual reports, Ministry of Health & Family Welfare, Government of India | The Rural Health Statistics (RHS) are the Government of India's annual account of public-health infrastructure and workforce at the sub-national level — Sub Centres (SCs), Primary Health Centres (PHCs), Community Health Centres (CHCs), the health workforce staffing them, building status, and separate tables for tribal areas. This project compiles 19 years of those reports into one consistent, analysis-ready dataset. ## 2. Files | File | Shape | Contents | |---|---|---| | `RHS_All_Years_Long.csv` | long | `Year, State, Category, Metric, Value` — the master table | | `RHS_Combined_Data_Dictionary.csv` | reference | all 863 indicators: code, description, domain, year coverage, raw source names | | `RHS_per_year_tables.zip` | wide | 19 CSVs, one per year, each State × Indicator as published | | `RHS_thematic_tables.zip` | wide | 12 domain tables, State × (Indicator-Year) | ## 3. How to read an indicator code Most codes follow the pattern **`[CADRE/FACILITY]_[AREA]_[STATUS]`**. - **Area:** `_R_` rural · `_U_` urban · `_T_` tribal. - **Status:** `_R` required (per IPHS norms) · `_S` sanctioned · `_P` in position · `_V` vacant · `_SH` shortfall. - **Facility counts:** `R_SC`/`P_SC`/`SH_SC` (required / in-position / shortfall Sub Centres); `F_SC_R` functional rural SCs (2019+ naming). - **Building position:** `_GB` government building · `_RB` rented · `_RFP` rent-free panchayat/voluntary · `_BRC` buildings required to be constructed · `_BUC` buildings under construction. Examples: `DOC_PHC_R_P` = doctors in position at rural PHCs; `SC_T_R` = required Sub Centres in tribal areas; `TSP_CHC_R_SH` = shortfall of total specialists at rural CHCs. The data dictionary gives a plain-English description for every code. ## 4. Special values — never zeroed Source markers are preserved exactly as reported and are **not** converted to zero: | Marker | Meaning | |---|---| | `*` | Surplus — in-position meets or exceeds requirement | | `**` | Surplus (alternate marker used in some tables) | | `NA` | Not available — the state did not report the figure | | `N App` | Not applicable — the facility type does not apply to that state/UT | In the long-format table these appear as the value or as an absent row (the long table carries numeric values; a `*`/`NA`/`N App` cell is not emitted as a numeric observation). A blank is meaningful: it is *not* a zero. ## 5. Two reporting eras (harmonisation) The RHS report was redesigned in 2019, changing many column names and adding Health & Wellness Centre (HWC) tables. This dataset maps **1,010 raw column names → 863 canonical indicators** so a single indicator can be followed across the break: - **2005–2018 (Era 1):** e.g. `DOC_PHC_S`, `F_SC`. - **2019–2023 (Era 2):** e.g. `DOC_PHC_R_S`, `F_SC_R`, plus HWC-integrated names. **2019 caveat.** In the 2019 round, Sub Centre and PHC counts were published under HWC-integrated column names (e.g. `SC_HWCSC_T_P`) reflecting the integration of upgraded facilities. These are the same underlying facilities; the dataset and the Explorer read those columns for 2019. Their values slot smoothly between 2018 and 2020. ## 6. Requirement basis (IPHS) "Required" figures follow the Indian Public Health Standards (IPHS) population norms (broadly: one SC per ~3,000–5,000 people, one PHC per ~20,000–30,000, one CHC per ~80,000–1,20,000, varying by rural/tribal/hilly area). Requirements are recalculated as census population estimates are updated, so the required counts step up at census transitions rather than drifting yearly. The 2005 report prints two requirement tables (1991- and 2001-census basis); this dataset uses the **2001-census** table for internal consistency with the following years. ## 7. Boundary and entity changes (tracked explicitly) - **Andhra Pradesh / Telangana:** Telangana was formed from Andhra Pradesh in 2014. Rows before 2015 are the **undivided** Andhra Pradesh. - **Jammu & Kashmir / Ladakh:** reorganised in 2019-20. J&K figures before 2020 include present-day Ladakh; Ladakh appears separately from 2020. - **Dadra & Nagar Haveli and Daman & Diu:** the two UTs merged in 2020. This dataset presents them **combined across all years** (summed before the merger) so the entity is continuous. - All 36 current States/UTs report across the series where the source does; historical entities bring the distinct-label count to 40. ## 8. Provenance and verification Built PDF-first: the printed report is treated as ground truth. 1. **Source** — 19 annual RHS reports (2005 to 2022-23) as released in PDF; where a year circulated only as a scanned/image PDF (2012, 2022-23), the Ministry's own spreadsheet annexes were used. 2. **Extraction** — tables parsed per report; OCR artefacts corrected against the printed page. 3. **Harmonisation** — raw columns mapped to canonical indicators; state names and boundary changes reconciled. 4. **Verification** — 1,300+ automated checks including structural integrity, arithmetic identities, cross-year consistency, and **value-by-value comparison of headline facility figures against the source tables for all 19 years**. Endpoint tables (2011, 2021) were verified 100%. ## 9. Changelog **v1.0 (2026)** — first public release. Includes three source-corrections made during verification and confirmed against the original reports: - **2006** — three consecutive workforce tables (Pharmacists / Lab Technicians / Nursing staff) were rotated in an earlier extraction; re-extracted from the source PDF (e.g. nursing in-position corrected to 28,930). - **2015** — every workforce table was shifted one column in an earlier extraction; re-extracted from the Ministry spreadsheets (e.g. doctors in position corrected to 27,421), and 75 previously-empty tribal workforce columns were filled. - **2022-23** — OCR-damaged state-name cells had dropped several rows from the digital annexes; restored via verified alias mapping, including the full Nagaland column and a blank-name Chhattisgarh row. ## 10. Known limitations - **Granularity** — state/UT level, not district. Sub-state variation is not captured. - **Presence, not quality** — a facility counted as functional is not necessarily adequately staffed, equipped, or utilised. - **Reporting gaps** — some state-years are reported "Not Applicable" or "Not Available" in the source; these are preserved as such, not imputed. - **Boundary transitions** — reorganisations (Telangana, Ladakh) may create step changes in a state's series independent of real infrastructure change. - **No imputation** — missing values are left missing throughout. ## 11. Access The dataset files are shared on request while the collection is prepared for formal release. Request access via the Explorer or by email (surajbhor.lkml@gmail.com) with a line on intended use. This documentation, the indicator dictionary, and the interactive Explorer are openly available. ## 12. License and citation Licensed **CC BY-NC 4.0** (Attribution-NonCommercial). See `LICENSE`. > Association for Socially Applicable Research (ASAR). *Rural Health Statistics > of India — harmonised dataset, 2005–2023.* Compiled from Ministry of Health & > Family Welfare, Rural Health Statistics annual reports. Version 1.0, 2026. > https://asarforindia.github.io/rhs-india/ This is an independent research compilation, not an official publication of the Ministry of Health & Family Welfare. For policy or academic use, cross-check figures against the original report of the relevant year.