Multisource Feedback and Systematic Bias

 

Multisource Feedback is currently an established method to obtain information on individuals at workplace. It is an essential component to be able to continue careers for doctors in the UK as it is a requirement for revalidation to remain on the medical register.

On colleague MSF a GMC report states:

‘There was evidence of systematic bias amongst some groups of colleague respondents in respect of assessments provided’

‘Certain doctor characteristics appeared to predict systematic variation in colleague assessments.’

‘In interpreting data arising from such surveys of colleagues, consideration should be given to the possibility of systematic bias in a colleague’s report based on non-clinical aspects of care, as well as the extent of colleagues’ familiarity with the doctor.’

 

On Patient MSF a GMC report states: 

'There was evidence of systematic bias in reporting amongst some groups of patient
respondents. '

'In interpreting data arising from such surveys of patients, consideration should
be given to the possibility of systematic bias in a patient’s report based on nonclinical aspects of care, as well as the socio-demographic profile of the patient
sample'

Software systems that are used for MSF have begun to use benchmarking – it is not clear if these benchmarks are statistically adjusted or standardized to consider and recalibrate for these biases. It is also not known whether the software systems identify these biases and inform the doctor that the responses could be impeded by biases.

It is not clear if doctors undergoing revalidation for GMC licence to practice are explicitly informed of these biases and offered guidance and support on how to choose MSF respondents to negate the existing biases or counselling on how to recognise, challenge and deal with any biases that becomes apparent.

Appraiser training does have an EDI component, but it is unsure if this training includes methods on how to consider these biases including any undue positive biases?

There are publications that suggest that the MSF could be one of the triggers that may be used when considering performance. This makes any issues regarding bias paramount to the topic.

It is very important to prevent issues – unfortunately it may be the case that the MSF methodology as currently implemented for doctors (and perhaps other health professionals) has a built-in flaw in the structure when it does not consider and adjust for biases.

Doctors should be aware of this and be prepared to address this issue should it affect them. Doctors’ organisations should resolve this at the earliest opportunity.

 

 

 

Ref:

  https://pubmed.ncbi.nlm.nih.gov/23095930/ 

 https://www.gmc-uk.org/-/media/gmc-site/about/rtguidanceresearchandpilotingofgmcquestionnairesdc8593pdf64915124.pdf

My bias declared publicly. Could you declare yours?

For quite sometime I have been active in quality improvement, leadership, patient safety, human factors and related areas. My interest includes discrimination in general but healthcare in particular and especially doctors. This relates to discrimination in the western world particularly in the UK. We know bias plays a part in this discrimination. Government and organisations have taken efforts in dealing with discrimination, this takes the form of law, training in equality and diversity and monitoring selected data. We are told how to deal with discrimination and how to avoid discrimination.

We were never asked to assess, quantify and declare our bias.

Bias is a part of life. For better or worse our history, legacy and as humans our capacity to pattern match means that bias gets in. We could even argue that bias is a normal part of life. However when bias affects the career, livelihood, reputation and freedom of certain segments of the population, it ceases to be normal. While bias could be accepted as normal, the positive or negative effects of bias that affects segments of population repeatedly can never be accepted as normal.

How do we prevent bias from affecting others?

In my opinion unless we assess our bias by a common method and then declare it publicly we cannot begin to tackle discrimination.

Once our bias is published publicly, then we can start having methods to take that into account in our decision making by appropriate counterbalances, algorithms, adjustments etc.

I cannot ask others to publicly declare their bias unless I have done so myself.

Currently, the Harvard Implicit Association Tests are a good (but not totally proven) method to assess our bias. https://implicit.harvard.edu/implicit/takeatest.html

I have taken the test that relates to race and I am posting the results by screenshot here.

I am aware that this test relates better to the USA and refers to the black-white races. I also took this test thrice; twice the same result came up that I had a moderate automatic preference for white people and once the test showed no preference between white and black people; on the occasion that the test showed no preference, I had deliberately slowed down to take the test.

So does that make me a racist? I do not think so. Does that make me biased? May be it does, but since I have known this for sometime now, I make deliberate conscious effort to overcome my Harvard IAT data based moderate automatic preference for white people. I think very deeply on how my actions are impacting. I take extra effort to review decisions before implementing them with the awareness of my bias to ensure that my actions do not affect black and minority ethnic people adversely.

If I was an examiner, investigator, police, judge, etc I would want my decisions to be counterbalanced by a formal algorithm to prevent my bias from hurting BME people.

I am not in an unusual category as far as my bias is concerned; I am part of the largest category of persons who have taken this test. How does this affect the society?

We need to start somewhere.

I encourage you to take the test and post it publicly. I am happy to add your test result to this blog if you send it to me. Let us make a change. Let us reduce bias. Let us take the first real step by publicly acknowledging our own bias.

Here is the list of persons who have agreed to go public with their bias tests. My sincere thanks for permitting me to add their names:

Dr Kim Holt: 'slight automatic preference to black people over white people'

Dr Joe Karthikapallil (Ophthalmologist) : 'moderate automatic preference to white people over black people'.

Dr Joydeep Grover: 'no automatic preference between African Americans and European Americans'

Dr Vivek Chhabra (ED): 'no automatic preference between black people and white people'

Mr Bhavik Patel (Financial Sector) - ' slight bias towards white individuals compared to black individuals'

©M HEMADRI

Follow me M HEMADRI on Twitter @HemadriTweets

M Hemadri’s mini e-book 'Standardised Management Conversation' is available - click http://www.amazon.co.uk/Standardised-Management-Conversation-Hemadri-ebook/dp/B018AWBJTU 

The Evolution of Language and its Impact on Racism

The Evolution of Language and its Impact on Racism

Some normal innocent original words took new meanings and won’t go away. It blights us today.

A muse on the four words, Fair, Black, Dark and White
(All etymology used in this essay is from Online Etymology Dictionary http://www.etymonline.com )

FAIR

Whenever we see injustice, someone or a group of persons negatively affected due to no fault of theirs we claim it is unfair, we then want to fight for fairness and we want a fair society.

Fair in this context means equitable and free from bias.

Let us look at the evolutions of the word ‘fair’. 

Originally the old original English meaning was ‘beautiful’ 'pleasing to sight, attractive' and similar; it also referred to good weather.  Then around the 1200 the word fair included references to 'light skinned people'. Are you beginning to see where we are going with this? The gradual slow association between pleasing to sight, attractiveness to lighter skin. Then by 1300 the word fair begins to be associated with "according with propriety; according with justice," and also refers good/auspicious and by 1400s it refers to "equitable, impartial, just, free from bias". From 1860 it refers to "comparatively good."

So the history of the word fair starts with beautiful which gets linked to light skinned persons, begins to refer also to justice and then good. 

Fair - white skin – just/unbiased – good 

That is the sequence associations of the evolution of the word fair.

DARK 

Let us now look at the word dark
The online Etymology dictionary says ‘Old English deorc "dark, obscure, gloomy; sad, cheerless; sinister, wicked," from Proto-Germanic *derkaz (cognates: Old High German tarchanjan "to hide, conceal"). "Absence of light" especially at night. Another old English link was darkly which meant horrible or foul.

That is the original meaning.

It is only as late as 1670 that dark begins to refer to ignorant. By 1700s it refers to negativity and by 1775 the word is used to refer to black people.  

The sequence here is
Dark/obscure/gloomy/sinister  -  Ignorant  -  Black people (and brown people)


BLACK

Original English had two meanings for the word Black. 'When used as a noun it referred to the colour Black. The online Etymology Dictionary says ‘When used as an adjective ''Old English blæc "dark," from Proto-Germanic *blakaz "burned" (cognates: Old Norse blakkr "dark," Old High German blah "black," Swedish bläck "ink," Dutch blaken "to burn"), from PIE *bhleg- "to burn, gleam, shine, flash" (cognates: Greek phlegein "to burn, scorch," Latin flagrare "to blaze, glow, burn"), from root *bhel- (1) "to shine, flash, burn;" Or bleach''. This was probably the reference to and origins of 'blacksmith' 

'The same root produced Old English blac "bright, shining, glittering, pale;" the connecting notions being, perhaps, "fire" (bright) and "burned" (dark). " Used of dark-skinned people in Old English.'

In 1300 it begins to refer to colour of sin and sorrow, by 1500 the word black began referring  to 'dark purposes and malignant' by 1500 it is also used to refer to dishonour, besmirch;  It then gets a reference to blackmail (perhaps because it burned and scorched and not a reference to colour). So we can see the links and connotations are changing and evolving from black being a reference to colour to references to negative and bad things.

The pnemonic and bubonic plague of the 1300s was at that time not called 'black death'. In the 16th century, after the 1300 to 1500 evolution of black from reference purely to colour or burn/blaze/glow/livid,  to include negative meanings, the word black was retrospectively attached to those tragic deaths.

By 1600 you have blacklists, by 1800 you have blacksheep

Black eye "discoloration around the eye from injury"  in 1600, which is obviously a description of a physical condition, morphs into  ‘Figurative sense of "injury to pride, rebuff"  by 1744. By 1880s it also means "bad reputation".

The term Blackguard from a jovial mocking reference in 1500s becomes a references to criminal classes by 1700s

By 1920s 'in the black' means a corporate loss and by 1930s references to 'black markets' are established.

Even the black widow spider has brown, red, black and yellow on its shell but gets called 'black'

The sequential evolution is from the original black which was a colour and a reference to black persons, and had very positive terms to illustrate it gradually moved over a long period of time to refer to everything bad and negative.



WHITE

The Online Etymology Dictionary says
‘White means …. "whiteness, white food, white of an egg," . Also in late Old English "a highly luminous color devoid of chroma. In old English the word White also referred to  "bright, radiant; clear, fair,"  In other language derivatives it meant  "white; to shine" (cognates: Sanskrit svetah "white;" Old Church Slavonic sviteti "to shine," svetu "light;" Lithuanian šviesti "to shine," svaityti "to brighten"). '

White was also used as a surname for persons with fair hair or complexion. 
It is indeed interesting that both black and white originally referred to 'shine' but obviously referring to different colours the meanings diverge as time passes by.

It is not as though the word white is not associated with negative meanings. But whenever it is associated with a negative meaning it is in comparison with another negativity where white comes out better off.


White collar crime – is associated with a softer version of behaviour, associated with genteel persons, linked with notions of education

White trash - is not when white people behaved in a trashy manner, it is when white people were in comparable positions to black servants (in America)

White Elephant - was a gift of honour though it was one which ruined you. The ‘elephant’ (i.e. the gifted object) is obviously of a high value, the maintenance of which cannot be afforded by an economically weaker person.

White wash - is cover up negativity, bad news, bad things.

White lie - is even a lie becomes a good thing; it is seen as a thing done to ‘benefit’ others.

	FAIR	DARK	BLACK	WHITE
Original meaning	Beautiful, pleasing, attractive	Obscure, gloomy, sad, wicked, absence of light	Colour Black, dark, burned, gleam, shine, flash, blaze, glow, bright, glittery, Also used for dark skinned people	Colour White, radiant, clear, fair, shine, brighten One with fair hair or complexion
1200	Original + light skinned people
1300	Both of above + Just/impartial		Sorrow, sin
1500s			Malignant purpose Dishonour, besmirch, blackmail Blackguard (jovial reference)
1600s		Ignorant	Blacklists, Blackeye (as in physical description)
1700s		Black people	Blackeye (as in Injury to pride) Blackguard (criminal class)
1800	All of above + ‘Comparatively good’		Blackeye (as in bad reputation)	White trash White elephant
1900s			In the black (as in financial loss) Black market	White wash White lie White collar crime

THE IMPACT OF THESE WORDS ON RACISM

Are you seeing how it works? Humans are pattern matchers. Human brain is not primarily for analysis though very capable of it. The default position is pattern matching. Here is a more recent example of pattern matching related distortion. The nazis used the swastika and now though the swastika is actually a powerful holy positive symbol used (potentially) by a billion people in other parts of the world, in the white western European-American world it is a negative symbol. We do not look at the meaning, we look at the usage and assign meaning. Our pattern matching in the western world, now fears the swastika as a nazi thing not how it actually is which is a Vedic thing.

Similarly our pattern matching is now helpless when it comes to racism, it associates black and dark with everything negative. It associates white with everything positive. We have becomes inadvertent slaves of aspects of our possibly distorted evolution of the English language. It served the purposes of white Europe in previous times.

In 1500s many European countries abolished slavery in their own countries as they were actively taking slaves from Africa. The language associations described above show a change in usage mostly in the 1500s and then it surges forward. This links clearly with the origins of the history of European involvement in the slavery of black people around that time and its growth. There seems to have been a desperate psychological need to create these new negative links on the color-people-trait theme so as to distance justify and garner support for the white Europeans who were trading black people as slaves. The links with dark colour and negativity has left a legacy which is now seemingly impossible to resolve.

Slavery itself was much older but that was not linked to colour. It was linked to debt, war, crime, punishment. We still see the effects of those kinds of linkages in human trafficking, poverty traps etc but those are linked to numericals, philosophy, economics; and not linguistic or visual based. Therefore when we think of say human trafficking we are able to look at it in an analytical mode when process, numbers becomes solutions. When we talk about racism we become pattern matches and emotion becomes a defense. 


CHALLENGING AND CHANGING THE LINGUISTICS

It can be expected that anyone who uses this rationale or logic for a change of language is going to face at the least a push back and at worse a back lash. The importance of context when interpreting words or phrases is paramount. Context however is immediate to the circumstance; even historic context is shadowed by our immediate context. Our current circumstance and context is preloaded with centuries of evolution and thus the links are woven into the genetics of the evolution of language. If context is indeed everything, then the language developed for the context of the 16^th 17^th and 18^th centuries becomes incorrect for usage in the 21^st century. By using such language in the 21^st century we become automatic accomplices to perpetuation of archaic practices.
 
It is not a simplistic matter of political correctness, it is not moral policing, it is not thought control. It is a matter of looking clearly at the evidence and its impact. Language links up with imagination and imagery then becomes etched in our mind to be called up on every occasion we pattern match (which is pretty much all the time) – these are the origin of unconscious and subconscious bias. This is the reason why often even black people taking the Implicit Association Test on race seem to be biased towards white people.

Colour needs to be disentangled from its links to bias. This will neither be easy nor quick. There are possibly many ways of doing this. Language architects can be engaged to weed out negative connotations. This need not always be people, software can enable this effort. The positive connotations of the words currently used negatively could be increased. The latter would be a better approach but understandably both approaches would be needed.

At a personal level, every reader of this article is urged to stop using at least these four words, FAIR, DARK, BLACK and WHITE for anything other that in its original meanings. Words such as black should not be used for accountancy (‘in the black’) or the various negative connotations. Black should be used only to refer to colour. Similarly with other words. Anything other than the original meanings should be subject to extraordinarily vigorous challenge and such usage be eliminated. The future generations may have a ray of hope for an impartial world (not a fairer world) if language distortions are ironed out. 

It can be done. It is not too late.

©M HEMADRI 

Follow me on twitter @HemadriTweets

Post Script
The links between dark/black and negative is mainly a Europe influenced phenomenon. In the parts or the world now known as India an incarnation of God was called Krishna which literally means the 'dark/black one' - Krishna is even now the God for a billion people with all its positive connotations. However, due to the power of English across the world and the history of Britain in its influence across the world even in India the same links between colour and positive or negative traits apply though not directly to Krishna (yet). Of course, that part of the world has its own linguistic history based biases persisting till today.


References and search term links

http://www.etymonline.com/index.php?term=fair
http://www.etymonline.com/index.php?allowed_in_frame=0&search=dark&searchmode=none
http://www.etymonline.com/index.php?allowed_in_frame=0&search=black&searchmode=none
http://www.etymonline.com/index.php?allowed_in_frame=0&search=white&searchmode=none

NB: While this blog site is primarily about healthcare, I have an interest in issues such as bias, race and its impact on healthcare. I hope a better understanding these issues in the broader context, helps us deal with healthcare in a better manner.