Recently NHS England and NHS improvement published the results of the NHS staff survey 2020. This collated responses from employees over an eight week period in the autumn of 2020. The survey covered all NHS trusts including foundation, acute, and specialist hospital trusts as well as ambulance, community, and mental health trusts. 7.7% of respondents are medical or dental staff, and 28% of nursing staff responded. The results have revealed a worrying trend with 44% of staff reporting feeling unwell as a result of work-related stress, which has increased from 40.3%, and is the highest result in five years. 48.2% of respondents have reported being discriminated against on the basis of their ethnic background and 7.7% on the basis of their gender. Wellbeing and morale among those who work for the NHS is reportedly poor, but does the NHS staff survey tell us the whole story, or are we only scratching the surface of what is going on behind closed doors?

Considering any dataset requires a certain degree of critical appraisal. In research terms it is important to consider bias with one of the most important questions applying to surveys being—is our surveyed population representative? Does it serve its purpose to adequately and accurately inform us on the subject about which we are asking?

Firstly, there is a degree of non-response bias applicable to the NHS staff survey. Response rate of the survey has never been above 60% in the last four years and this year was only 49.1%. If only half of the workforce are returning their surveys, are these participants representative? Are there varying degrees of response that are proportional to how disenchanted a person is feeling? It could be argued that those people whose morale is particularly low may feel so disempowered that they simply feel there is no point expressing their viewpoint because no one will listen to, care about, or action their concerns. 

Secondary we need to ensure that the responses we are getting are representative of all groups in order to avoid selection bias. Primarily this relies on ensuring groups are adequately represented in the respondent population, but it can also rely on asking the right people the right question. Ask a predominantly White British department if they think there is a problem with racism and it is likely the answer will be no. Ask the one worker from an ethnic minority group, who is more likely to have been subject to said discrimination whether they have experienced racism or discrimination in the same department, and you may well find that their answer is different. Notable also are those that are not surveyed—those on long-term sickness or who are bank staff. They will undoubtedly have a unique view, particularly if their sickness is work related, that we are missing out on. 

The grouping of people into white vs ethnic minority can be harmfully reductionist in this setting. Looking closely at the 2020 survey data while those who are from a White British ethnic background scored their workplace’s equality, diversity, and inclusion as 9.3/10, those from a Black Caribbean background scored only 7.2/10. However, those who identify as being from a Chinese ethnic background scored EDI (equality, diversity, inclusion) 8.6/10. Not all minority ethnic groups are experiencing the same barriers to EDI in a ubiquitous way and reading the headlines of the data without examining the minutiae can miss potential opportunities to enhance positive experience of the workplace.  

When interrogating the opinion of any large group of people, applying a blanket approach to gauge the opinion of a workforce can not only miss the nuance of human experience due to its lack of granularity, but can also be falsely reassuring if you fail to ask the people most likely to be negatively affected. The NHS staff survey tells us something, but it does not tell us everything.

While trusts should be allowed to celebrate results showing “excellent satisfaction” it is imperative that we dig deep and interrogate the data closely to ensure that we are not seduced by positive data that may disguise more insidious and harmful issues. In a year where staff have been faced with a volley of covid-19 related difficulties both in and out of work, it is hardly surprising that wellbeing and morale are low, but the question now is how we fix that. The key to unlocking this may well lie in the details of these data. If we want to fix issues with institutionalised racism, sexism, bullying, or harassment within the NHS we need to ask the right people the right questions. This is not always comfortable or the easy path, but by having the courage to ask the questions we might not always like the answer to and accepting that these results only tell us part of the story we can to continue to improve and ensure that NHS staff feel safe and engaged in their workplace. 

Clara Munro, editorial registrar and clinical fellow, The BMJ, and general surgical trainee, North East England.

Competing interests: None declared.