Clare Griffiths
LightNews LightNews
banner
statsgeekclare.bsky.social
Clare Griffiths
@statsgeekclare.bsky.social
9.6K followers 1.1K following 2.6K posts
Demographer, former tweeting statistician and female data lad. Attempting to make a difference in public health analysis at DHSC. I ran the UK's covid dashboard. Now seen running parkruns. DMs not available. She/her 🍉🏳️‍⚧️🏳️‍🌈
Posts Replies Media Videos
statsgeekclare.bsky.social
Especially when you consider this (from recent CMO report) www.gov.uk/government/p...
Women under 30 have rapidly falling cervical cancer rates due to human papillomavirus (HPV) vaccination which was introduced in 2008. Trend in cervical cancer incidence by age group (age standardised rates per 100,000 population), England, 2001 to 2022. In England, the HPV vaccine has been offered to girls in school year 8 since September 2008. Source: NHS England, Cancer Registration Statistics, England, 2022 (2025 data refresh) as at 12 June 2025, Cancer Registration Statistics, England 2022 supplementary request.
November 2, 2025 at 8:00 PM
statsgeekclare.bsky.social
Rounded off @royalstatsoc.bsky.social conference week with an obligatory parkrun at Oriam! Thanks @fjtweedie.bsky.social for enabling the visit and marshalling. It’s a lovely parkrun, nice trails, a few out and backs, friendly team! #LoveParkrun
Me and Fiona after the run Oriam sports centre Oriam parkrun pop up sign having a lie down on the grass A woodland trail path with trees either side
September 6, 2025 at 1:09 PM
statsgeekclare.bsky.social
Lovely evening at the RSS Conference dinner after a fabulous week in Edinburgh. Bring on Bournemouth next year! @statsrhian.bsky.social @scottheald72.bsky.social @altealorenzo.bsky.social @fedeandreis.bsky.social
Me, Altea and Kim at the RSS Conference dinner Me Scott and Altea at the RSS Conference dinner Me and Fede at the RSS Conference dinner Me and Rhian at the RSS Conference dinner
September 5, 2025 at 12:39 AM
statsgeekclare.bsky.social
Interesting session at the RSS Conference on a systems approach to collaboration with some really useful points on how resource constraints reduce collaboration and cause organisations to focus inwards. @royalstatsoc.bsky.social
PowerPoint slide in a lecture hall with speaker standing to one side: Title: Organisational Funding and Autonomy › Funding cuts to local government to alleviate funding pressures › To alleviate knowledge gaps and capacity constraints, external consultants deployed: • Increasing resource constraints • Reducing in-house capacity building › Where consultants' lack contextual knowledge, may reduce the usefulness of analysis for decision-making. This affects decision-making quality, increasing resource constraints
September 4, 2025 at 8:36 AM
statsgeekclare.bsky.social
On Wednesdays we use log scales @royalstatsoc.bsky.social
A woman with grey hair and glasses holding an RSS lanyard containing three badges reading DATA IS THE NEW PUNK I AM A SAD DECISION-OOPS! STATISTICIAN #GENERATIVEAI On Wednesdays we use log scales
September 3, 2025 at 3:30 PM
statsgeekclare.bsky.social
Morning! First day proper of the RSS conference and I am starting with positivity.
A PowerPoint slide entitled Unleashing the Data Dividend Dr Paula McLeod and Dr Nicci Potts UK Foreign, Commonwealth and Development Office RSS International Conference 2025 In front of this is a lecture hall of seats with people’s heads.
September 2, 2025 at 8:12 AM
statsgeekclare.bsky.social
On my way to @royalstatsoc.bsky.social international conference! Being good and going by train instead of flying. Come to my session on accessibility in official statistics on Thursday, there’s a quiz!
Blue Lumo train in platform 3 at London Kings Cross station. Sunlight streams through the roof.
September 1, 2025 at 11:13 AM
statsgeekclare.bsky.social
Going through some family albums and found this pic from the Festival of Britain, amongst others!
A black and white photo taken in 1951 of the Skylon at the Festival of Britain.
June 20, 2025 at 7:06 PM
statsgeekclare.bsky.social
A couple of charts from our 20th century trends article, would love to find time to update it!
Charts showing trends in mortality from 1911 to 2000 in England and Wales, for major categories of disease and males and females separately. Discontinuity caused by coding changes stand out in places. Trends in heart diseases and stroke 1950-2000 showing how trends in ischaemic heart disease depend on whether other myocardial degeneration is included or not.
February 18, 2025 at 9:59 AM
statsgeekclare.bsky.social
R after a few years
January 21, 2025 at 9:54 AM
statsgeekclare.bsky.social
Well no surprises there! Parkrun time!
Strava year in sport graphic - top day of the week: Saturday, most active time: 9-10am
December 13, 2024 at 12:38 PM
statsgeekclare.bsky.social
Extremely proud of my team who have been shortlisted in the Analysis in Government awards for communication for our Health trends in England tool, designed to open up public health data to a broader audience. We used Quarto by @posit.co to build it. fingertips.phe.org.uk/static-repor...
On the left hand side - a screenshot of our health trends in England tool - showing six charts and some descriptive text. Below this is a picture of 6 people in a sports stadium. They are all (mostly) smiling. On the right hand side is a fancy picture of some tower blocks, an Analysis Function logo and the words - our work has been shortlisted! Analysis in Government Awards #AiGAwards
December 13, 2024 at 9:47 AM
statsgeekclare.bsky.social
Crikey it was a bit damp at parkrun this morning. Thanks to all the volunteers out in the rain and wind at Roding Valley!
A selfie of me, smiling, I am wearing a stripy hat, and green parkrun 250 t shirt.
November 23, 2024 at 11:55 AM
statsgeekclare.bsky.social
On my way to Leeds for a couple of days of work stuff and we all know how much I love Leeds 🤔. Still, at least I have coffee.
A cup of coffee on a train seat back table, window in the background showing train tracks
November 18, 2024 at 11:12 AM
statsgeekclare.bsky.social
Table 1
Table A tactore; ICD chapters and groups of causes with more than one thousand deaths. Aged 0-74 All ages Cause of death Number Number Conver- Conver- Number nunter factor* factor* 2,295 0.902 1,371 0.949 INTECTIOUS AND PARASITIC DISEASES 924 140, 101 0.972 86,993 0.984 53,108 001-139 0.969 1,213 0.984 NSOPLASKS 1,763 140-239 MALIGNANT NEOPLASMS: - oral cavity and pharynx 41,112 0.973 23,116 0.983 550 240-149 150-259 digestive organs and peritoneun 0.989 0.991 17,996 160-165 respiratory and intrathoracie organs 37,155 0.963 10,548 11,610 170-175 connective tissue, skin and breast 15,664 0.988 21,148 0.962 12,083 0.983 5,116 9,065 genitourinary organs 12,797 0.974 7,893 0.983 190-199 other and unspecified sites 8,687 0.957 5,605 0.968 4,904 200-208 lymphatic and haenatopoletie tissue 3,082 240-279 IIE ENDOCRINE, NUTRITIONAL AND NETABOLIC 8,499 0.708 3,720 0.826 DISEASES AND INKUNITY DISORDERS 4,779 6,601 0.699 2,747 0.806 250-259 Diseases of other endocrine glands 3,854 270-279 other netabolie disorders and immunity 1,077 0.854 746 0.914 331 disorders 280-289 IV DISEASES OF BLOOD AND BLOOD-FORMING ORGANS 2,100 0.762 0.885 1,412 290-319 10,744 0.426 2,039 0.520 MENTAL DISORDERS 8,705 290-299 Organie psychotic conditions, other 10,009 0.410 1,535 0.435 8,474 320-389 DISEASES OF THE NERVOUS SYSTEM AND SENSE ORGANS 10,483 0.710 4,969 0.818 330-337 5,514 Bereditary and degenezative diseases of central nervous systen 6,035 0.675 2,313 0.794 3,722 Other disorders of the central nervous system 3,645 0.731 2,030 0.817 1,615 390-459 393-398 VII DISEASES OF CIRCULATORY SYSTEM 278,849 0.986 117,488 1.000 161,363 Acute rheunatic fever, heart disease chronic rheunatic 2,931 0.950 0.929 1,849 1,082 4,044 410-414 Hypertensive disease 0.981 2,110 0.970 0.952 157,506 415-417 Lschaenic heart arseasel 2,534 Diseases of pulsonary circulation 79,057 0.991 2,070 78,449 Other forns of heart disease 1.812 883 1.869 1,187 430-438 20,355 Cerebrovascular d1sease 1.176 4,739 1.202 15,616 Disease… Table A - continued ICD Cause of death number 111 ages Aged 0-74 Aged 75 and over Number Conver- sion factor* Number Conver= sion Lactor* Number Conver 740-759 XIV CONGENITAL ANOMALIES 3,017 0.950 2,839 0.951 Lactor* 760-779 XV CERTAIN CONDITIONS ORIGINATING IN THE PERINATAL PERIOD 2,289 0.997 2,289 0.997 0.921 780-799 XVI SIGNS, SYMPTOMS AND ILL-DEFINED CONDITIONS 797-799 Ill-defined and unknown causes of morbidity and mortality 2,534 2,480 1.000 1.000 1,401 1,383 1.000 1.000 1,133 1.000 1,097 1.000 E800-8999 XVII EXTERNAL CAUSES OF INJURY AND POISONING 18,835 0.998 13,918 E810-E819 Motor vehicle traffic accidents 5,012 1.000 0.999 4,917 0.993 E880-8888 Accidental falls 4,334 1.000 678 1.000 8950-£959 Suicide and self-inflicted injury 3,928 4,315 0.995 1.000 1,161 3,868 0.999 1.000 2,767 0.993 E980-£989 Injury undetermined whether accidentally 447 1.000 or purposely inflicted 1,563 0.999 1,396 0.999 167 1.000 * The number of deaths under the 1983 coding rules divided by the number of deaths under the 1984 coding rules? ix
October 21, 2024 at 3:11 PM
statsgeekclare.bsky.social
Preamble
Front cover of DH2 no 11 Introduction Mortality statistics: cause 1984 deals with statistics relating to England and Wales. Comparable statistics for earlier years and separate statistics for Scotland and Northern Ireland are published as follows: For England and Wales for 1974 onwards - in the booklet Mortality statistics: cause; for previous years - in the Registrar General's Statistical Review of England and Wales, Part I, Tables 1, 17 and 17a. For Scotland - in the Annual Report of the Registrar General for Scotland, Part I. For Northern Ireland - in the Annual Report of the Registrar General for Northern Ireland. A summary of mortality statistics for the United Kingdom and constituent countries appears in the Annual Abstract of Statistics issued by the Central Statistical Office. Changes introduced in 1984: coding of the underlying cause of death from Part II of the death certificate (WHO Rule 3) A new format death certificate was introduced in England and Wales in 1927; there were two parts to the certificate, with the first part split into Ia, Ib, and Ic. The intention was that the certifier would record the morbid events leading to death in Part I of the certificate and in Part II enter diseases that contributed to death but were not part of the direct sequence. In 1948 The World Health Organisation (WHO) released the 6th revision of the International Classification of Diseases, Injuries and Causes of Death (ICD). This recommended for international use a certificate modelled on that for England and Wales; in addition rules were provided to lead to consistent of the coding underlying cause of death. Rule 1 (d) stated that where the condition in Part I is clearly a direct sequel to the condition in Part Il, the latter is to be preferred as the underlying cause. With some change in wording the same rule appeared in the 7th revision of the ICD. In the 8th revision of the ICD this became Rule 3, "If the condition selected by the general rule or rules 1 or 2 can be considered a direct seq… certificates (excluding those with only one cause mentioned on the certificate, or deaths from violent causes) revealed 238 where a "terminal event" appeared in Part I of the certificate as the only cause of death, while a major disease appeared in Part II of the certificate recorded as a contributory factor. It was suggested from this that incorrect assignment of the underlying cause to terminal events may be occurring for an apprec-lable number of deaths, and that this particularly applied to deaths from bronchopneumonia. In May 1983 WHO held an expert committee meeting to discuss the application and revision of ICD coding rules. Following this meeting and in the light of the above information, it was decided to supplement the guidance to OPCS coders from 1 January 1984 in the use of the WHO Rule 3. Where there was major disease recorded in Part II of the death certificate, and the underlying cause recorded in Part I of the cert- ificate was classifiable to the following conditions, Rule 3 was to be implemented: - 4151 Pulmonary embolism 4275 Cardiac arrest 428 Heart Failure 4511 Phlebitis and thrombophlebitis of deep vessels of lower extremities 4512 Phlebitis and thrombo- unspecified phlebitis of lower extremities, 4519 Phlebitis and thrombo- 4539 phlebitis of unspecified site Venous embolism or thrombosis 485 486 514 of unspecified site Bronchopneumonia, unspecified Pneumonia, unspecified 5728 Pulmonary congestion or hypostasis Hepatic failure or liver failure In the initial stages of implementation of Rule 3, duplicate copies of the cause of death details were scrutinised in the WHO Centre at OPCS, and enquiries sent to a sample of certifiers to discuss the suitability of coding the underlying cause of death to the information provided in Part Initiale certificate The replies to an initial enquiry of the handling sy testhanaling after consideration certificates for cent of the sample of respondents agreed that the material the certificate in Part 11 g should b… со. Tiew was ite • 1, ch. 34 ime en- L1 -- La- oli- Le- vere ed red able ould es 984• rea njury. pter, If no entry is provided in Table 4 for a cause that is listed in Table 2, this indicates that there were no deaths for that rubric coded after application of Rule 3. There are some entries where there were no deaths coded under the 1983 rule, but at least one death using Rule 3. Table 4 is laid out with columns for all ages and two broad age-groups 0-74, and 75 and over. Within these columns the numbers of deaths according to the 1983 and 1984 rules are provided, together with a conversion factor which is the result of dividing the figure in column A by the figure in column B. These conversion factors can be used, with care, to indicate the numbers of deaths that might appear using the 1983 rules for the specific cause of interest. There is a problem with small numbers (many of the factors are based on less than 10 deaths); there is also an assumption that the factor does not vary within the broad age-groups utilised, or between the sexes. In order to provide a quick introduction to Table 4, the following abstract (Table A) has been prepared. This shows for individual chapters of the ICD, and main subgroups where at least one thousand deaths have occurr-ed, the number of deaths according to the Rule 3 code and the conversion factor. Again the material is for persons at all ages and for the two broad age-groups. There are a few chapters where the conversion factor is less than 0.9: III Endocrine and other diseases, IV Blood and blood forming organs, V Mental disorders, VI Diseases of the nervous system and sense organs, XII Diseases of the skin and subcutaneous tissue, XIII Diseases of the musculoskeletal system and connective tissue, The lowest conversion factor is for Mental disorders, 0.43. This indicates that there has been appreciable transfer of deaths to these chapters under the application of Rule 3. There is only one chapter where the conversion factor is g…
October 21, 2024 at 3:09 PM
statsgeekclare.bsky.social
Here’s the @royalstatsoc.bsky.social Statistics Under Pressure steering group with Ed Humpherson from the Office for Stats Regulation and two of our speakers, after our panel session at the recent RSS conference
A group of people standing on a stage in front of a sign with the RSS Brighton Conference logo. Pictured L to R: John Aston, Sarah Walker, Victoria Cox, Paul Allin, Ed Humpherson, Clare Griffiths, Christl Donnelly, Fran Bryden
September 23, 2024 at 2:22 PM
statsgeekclare.bsky.social
And this one!! @altealorenzo.bsky.social
Two heads, smiling, poking out of a board showing a body in a suit and a body in a dress
September 23, 2024 at 2:17 PM
statsgeekclare.bsky.social
@altealorenzo.bsky.social look at this lovely pic of us in not so sunny Brighton!
Altea and I standing on Brighton pier in the rain - behind us is the Brighton Palace Pier sign and we are both smiling despite the weather.
September 23, 2024 at 2:09 PM
statsgeekclare.bsky.social
Bluesky now has over 10 million users, and I was #609,047! Woo
September 18, 2024 at 8:35 AM
statsgeekclare.bsky.social
The three amigos of official stats @scottheald72.bsky.social and Jen not-on-Bluesky-yet 🤣 #RSS2024conf
Me, Scott and Jen
September 5, 2024 at 10:27 PM
statsgeekclare.bsky.social
An early situation centre created by Mary Spear!
A PERMANENT EXHIBIT OF CHARTS HAS BEEN ESTABLISHED IN CONFERENCE ROOM 6221 OF THE DEPARTMENT OF LABOR THE CHARTS WILL CURRENTLY RECORD, IN GRAPHIC FORM THE WORK OF THE BUREAU OF LABOR STATISTICS. INTERESTED PERSONS ARE CORDIALLY INVITED TO VISIT THE EXHIBIT AT ANY TIME WHEN CONFERENCE IS NOT IN SESSION. FOR FURTHER INFORMATION CALL MARY E. SPEAR EXECUTIVE 2420 EXTENSION 165 or 617 Didactics She had a passion for sharing her knowledge and promoting good practice. She lectured "Graphic presentation of Statistics" at the American University (1947 - 1957) "show 'em you can do it!" Dr Altea Lorenzo-Arribas, RSS Conference 2024
September 5, 2024 at 11:08 AM
statsgeekclare.bsky.social
@altealorenzo.bsky.social presenting on Mary Eleanor Spear and her contribution to data visualisation for the Hidden Figures session at #RSS2024conf, celebrating the contributions that don’t always get celebrated or taught
Altea at the podium in the main hall
September 5, 2024 at 11:05 AM
statsgeekclare.bsky.social
Is the English Premier League competitive? Find out with statistics. #RSS2024conf
Two slides with the title Is the English premier league competitive? Presented by Nial Friel – University College Dublin Data arising from sporting competitions provide an abundance of interesting statistical challenges, from both a modelling and an inferential perspective. This talk will explore some of these challenges, including the question of how to cluster and rank teams (or individuals) based on pairwise comparison data. That is, when the data consists of the outcome when pairs of teams (or individuals) play against one another. A particular example will illustrate a modelling framework that can be used to assess how competitive the English Premier League has been over the past 40 years.
September 5, 2024 at 9:33 AM
statsgeekclare.bsky.social
One thing I love about the RSS conference is getting to go to sessions that I would never normally get to hear about - I always like to go to the sports statistics ones as they are genuinely fascinating.
A slide with the following text: Title: Sporting Contest Design • Designing optimal sporting contests is difficult! • Sports leagues have multiple (possibly conflicting) objectives. • E.g., reverse order draft (Booth, 2005; Fornwagner, 2019; Gong et al., 2021). • Kendall & Lenten (2017) extend this in multiple other scenarios. Picture: Olympic badminton players disqualified for trying to lose
September 5, 2024 at 8:12 AM