This article has Open Peer Review reports available.
Is perception of quality more important than technical quality in patient video cases?
© Roland et al. 2015
Received: 5 May 2015
Accepted: 29 July 2015
Published: 13 August 2015
The use of video cases to demonstrate key signs and symptoms in patients (patient video cases or PVCs) is a rapidly expanding field. The aims of this study were to evaluate whether the technical quality, or judgement of quality, of a video clip influences a paediatrician’s judgment on acuity of the case and assess the relationship between perception of quality and the technical quality of a selection of video clips.
Participants (12 senior consultant paediatricians attending an examination workshop) individually categorised 28 PVCs into one of 3 possible acuities and then described the quality of the image seen. The PVCs had been converted into four different technical qualities (differing bit rates ranging from excellent to low quality).
Participants’ assessment of quality and the actual industry standard of the PVC were independent (333 distinct observations, spearmans rho = 0.0410, p = 0.4564). Agreement between actual acuity and participants’ judgement was generally good at higher acuities but moderate at medium/low acuities of illness (overall correlation 0.664). Perception of the quality of the clip was related to correct assignment of acuity regardless of the technical quality of the clip (number of obs = 330, z = 2.07, p = 0.038).
It is important to benchmark PVCs prior to use in learning resources as experts may not agree on the information within, or quality of, the clip. It appears, although PVCs may be beneficial in a pedagogical context, the perception of quality of clip may be an important determinant of an expert’s decision making.
The potential benefits of patient video cases (PVCs) are being increasingly realised , with a survey of children’s hospitals in North America and the UK finding that video recordings of clinical interactions and patient signs are relatively common . Video can be a powerful tool as the addition of video to audio clips have been demonstrated to have large effects on the recall of the content of the cases, both objectively and subjectively . The knowledge and learning obtained from PVCs is dependent on a number of factors that have yet to be determined. Information content, technical quality, monitor fidelity, bandwidth availability, processing speed (if digital recording) and interference from other electronic devices all may influence learning from Patient Video Cases .
Quality issues are clearly relevant to the validity of assessments of learning outcomes however there is no universal definition of what constitutes high quality video. In a medical context most investigation of video quality has been in relation to telemedicine. The focus of this research has been either in the transfer of single pictures (such as in tele-dermatology ), communication between healthcare professionals and patients separated by large geographic distances  or specific radiological examinations such as echocardiograms . The endpoint of these studies being a comparison between different clinicians or clinical outcomes of patients, with no examination of the effects of image quality. Literature in the area of image quality and decision making is scant. The Royal College of Paediatrics and Child Health have utilised video cases in their postgraduate examinations since 2004. A review of the process found this particular assessment to be as reliable and informative as more traditional examination methodologies . In 2005 McFaul  conducted a feasibility study to test whether real time video pictures used in a clinical environment could be transmitted within a hospital to enable a specialist to form a useful assessment of the severity and nature of a child’s illness. In approximately 70 % of cases the image was felt to be good enough to guide clinical management, although there was variation between particular conditions.
The aim of this study was to assess paediatricians’ perceptions of video quality and acuity of illness in a selection of PVCs of unwell children. The null hypothesis being the technical quality of a video case of an acutely unwell child does not influence a paediatrician’s judgement on the child’s acuity.
Assess the relationship between senior paediatricians’ perception of quality and the measured technical quality of a selection of video cases.
Define the correlation between the paediatricians’ judgments on the acuity of patients clinical signs via video cases and the actual patient acuity
Evaluate whether the technical quality, or judgement of quality, of a video case influences the paediatricians judgment on acuity.
Schemata for viewing videos (each group contained 3 paediatricians)
Respiratory low quality
Respiratory medium quality
Respiratory high quality
Respiratory excellent quality
Hydration medium quality
Hydration high quality
Hydration excellent quality
Hydration low quality
Response to social cues high quality
Response to social cues excellent quality
Response to social cues low quality
Response to social cues medium quality
The four qualities repeated through the five patient categories (Respiratory hydration, Response to social cues, State variation and Colour)
Colour excellent quality
Colour low quality
Colour medium quality
Colour high quality
Acuity scoring grid to assess clinical features seen in video clips (as used by McFaul )
Pale or Flushed or Mottled
Cyanotic or Ashen
Response to social overture
Chats or smiles OR “alerts” (< 2months)
Single words or briefly smiles OR “alerts” briefly (< 2months)
No smile. Face anxious OR dull and expressionless or no “alertness”
If awake stays awake OR if asleep and stimulated wakes quickly
Eyes close briefly and then awakens OR awakens after prolonged stimulation
Falls asleep when examined OR will not rouse
Skin normal, eyes normal and mucous membranes moist
Skin/eyes normal and mouth slightly dry
Skin doughy or tented and dry mucous membranes and/or sunken eyes
Some distress eg recession
Laboured with grunt or nasal flare OR marked recession OR absent resps
The principal investigator had directly observed the majority of the children in a clinical context as they were videoed, and made a clinical judgement of PVC quality against the McFaul system, which was used as the gold criterion standard. The scenario described to all participants was that they were making a telemedicine judgement based on a video clip shown to them by a member of junior medical staff. They were asked to grade each clip. If they would be comfortable making a clinical decision based on the clip then this would be graded as at least 4 (very good – safe for clinical practice). The participants all viewed the clips from the same angle on the same computer screen (an LCD screen with 1366x768 pixel frame) in the same lighting conditions. No clips were longer than 20 s and the entire time needed per participant to complete all the questions was approximately 25 min. Clip length was chosen pragmatically to clearly demonstrate the clinical sign in question and allow for a manageable amount of cases to be studied in the available time.
Given the novel nature of the methodology power calculations were not undertaken as there were no prior studies with data to enable an estimation of effect size. Stata version 13 was used to analyse data with significance set at p < 0.05. Spearman’s rho was used to determine relationships between image quality standards and raters’ quality of assessment. This non-parametric test was chosen as the results were not normally distributed. In order to assess whether the level of significance of the correlation coefficient was affected by the clustering of ratings within the same video clips, corresponding tests were carried out using linear regression with random effects for the clips. Spearman’s rank correlation was used to determine the relationship between gold standard assessment and the raters’ assessment.
Twelve participants undertook the study. The total number of clips with quality scores was 333 (27 clips in 4 different versions seen by 3 participants at each version and 1 clip seen in 3 different versions by 3 participants at each version). There were three instances where a participant (three different individuals) was unable to make an acuity judgement on the clip. So there were 330 responses with acuity and quality scores for analysis.
Image quality versus Paediatrician assessment of quality
Quality of image
Paediatrician assessment of quality of image
Match of the paediatricians’ acuity score with the gold standard
Gold standard severity of patient
Paediatricians’ acuity score (%)
Paediatricians’ acuity score versus gold standard across the domains
+/− 1 difference
+/− 2 difference
84 (79.3 %)
20 (18.8 %)
2 (1.9 %)
45 (75.0 %)
12 (20.0 %)
3 (5.0 %)
Response to social overture
42 (70.0 %)
11 (18.3 %)
7 (11.7 %)
29 (61.7 %)
17 (36.2 %)
1 (2.1 %)
22 (38.6 %)
0 (0 %)
Technical quality of image versus acuity score
Industry defined technical quality of image
22 (26.2 %)
62 (73.8 %)
23 (28.4 %)
58 (71.6 %)
27 (32.1 %)
57 (67.9 %)
26 (30.9 %)
58 (69.1 %)
98 (29.4 %)
235 (70.6 %)
Although quality assurance of videos in respect of descriptive information content has been defined this has not occurred for video image quality . In this study the null hypothesis that the technical quality of a video clip of an acutely unwell child does not influence a paediatrician’s judgement on their acuity was accepted. However the paediatrician’s perception of the quality of the clip did appear to have an effect. It also appeared that judgements between paediatricians were generally cohesive for higher acuity patients but there was greater variation in more mild/moderate illness (i.e., lower acuity illness).
These results imply that the clinicians had a different meaning of “quality” from the technical industry standards. In other words the technical video quality measures do not match the factors which the paediatricians perceived as important in being able to judge acuity in PVCs. Further studies are needed to elucidate the factors underlying a clinician’s perception of video quality. The reasons why the paediatricians may be influenced by their perception of quality rather than the technical quality are complex. Polanyi  is credited with coining the term ‘tacit knowledge’, a concept Schön  described as knowledge that is usable but which one can not rationally express. Given that all the paediatricians were given the same visual information and, lack of specific clinical details, the intrinsic cognitive load  for all was the same. In the absence of a qualitative analysis, it is difficult to know whether experience, clinical knowledge or tacit knowledge contributed to the decisions the consultants made. It may be that the clips rated as poor quality were the cases where there existed the greatest discrepancy between the clinical sign shown and need for further information to evaluate those signs. A young infant with a high respiratory rate and a background of prematurity has a greater risk of subsequent deterioration. This information is normally vital in order to make a decision about care. Although the consultants were not being asked about disposition or treatment they may have struggled making a decision without this type of context.
Context has shown to be very important in decision-making. Croskerry cites examples of system 1 and 2 processing, a psychological theory regarding cognitive reasoning . Sherbino et al.  described system 1 as rapid, unconscious, and contextual thinking whereas system 2 is slow, logical, and rational. Kahneman and Tversky  put forward the original theory that system 1 thinking results in error when system 2 processes do not spot mistakes during system 1 processing. The consultants when reviewing real patients would be making instinctive judgements, likely with background information readily available or already processed. The insufficient information in this study meaning system 1 decisions were difficult to process and there was not sufficient time in the study to utilise system 2 analysis. Although the system 1 and 2 classification itself has been questioned , it is plausible that the quick decisions reached when viewing video footage must be informed by supplementary information. Video assessment may not be authentic without this context because the second phase of cognitive evaluation may not be possible if the initial information is insufficient (clinicians may be concentrating on what is missing rather than what they see).
It is possible that certain clinical signs were more difficult to interpret than others and this may have influenced the relationship between video quality and outcome. There is some evidence for this as there was a greater proportion of agreement with the gold standard for the domains of alertness (75.0 %) and colour (79.3 %) than there was for hydration (61.7 %) or respiratory (61.4 %). This could be considered surprising as colour gradation is affected by the quality of the image and has poor inter-observer reliability even when examined in direct clinical practice . The fact that colour was not commonly incorrectly assigned, low incidence of +1 (18.9 %) and +2 (1.9 %) differences compared to the average +1 (24.9 %) and +2 (3.9 %) difference with gold standard acuity lends credence to the fact that it may not have been the technical quality the participants were judging but an overall information deficit. Ideally a range of acuities in the same patient (or at least similar ages and ethnicities) across a range of qualities could be utilised to try and limit the impact of these potential contextual confounders.
In order to confirm these findings a validation of this study is needed as a number of factors may have influenced the results. The participants may not be representative of typical clinicians as they were all experienced in assessment. It could also be argued that the gold standard was not adequate as less than 70 % of the participants were correct for hydration and respiratory. However, even if the gold standard was accepted to be incorrect, it would still be the case there was large variation in the answers. The purpose of the study was to benchmark quality and observe variability between experts rather than test clinical accuracy, so this pragmatic gold standard was thought to be sufficient. Similarly other measures of video quality could have been varied, such as signal-to-noise ratios, but bitrates were practically the easiest to alter with available resources. However it would be useful to replicate the study using a reference point video as a control to confirm the clinical significance of the bitrates used. A further study altering clip length may also be useful as this may have impacted on the judgement of the paediatrician and it is possible showing the clips for a longer period may have altered the acuity and quality score given.
Preference for a particular cinematic film compared to another has been shown to affect the perceived quality of the video . It is unlikely the paediatricians will have had a favourite video case which may have biased their judgement as these are scenarios they would associate with a need to make a clinical rather than emotional decision. However had an initial study examining the factors influencing clinician’s perception of quality been performed it is possible it may have revealed design features to control for in this research project.
A recent study has demonstrated considerable variability in the assessment of breathing difficulty in wheezy children between professionals . Utilising video to educate and quality assure could become increasingly important and the outcomes of this work may also have application beyond Medical Education as minimum standards would be useful to aid telemedicine service providers to configure data transfer links appropriately.
It is important to benchmark PVCs prior to use in learning resources as experts may not agree on the information within or quality of clip. Although PVCs may be beneficial in a pedagogical context their translation into an assessment methodology or clinical use via telemedicine must be further examined as perception of quality of clip was an important determinant of an expert’s decision-making.
The authors acknowledge Dr. Nic Blackwell of OCB Media in the conversion of the video cases into different bitrates. The authors are also grateful to the examinations committee of the RCPCH for allowing the data collection to take place.
This report is independent research arising from a Doctoral Research Fellowship supported by the National Institute for Health Research. The views expressed in this publication are those of the author(s) and not necessarily those of the NHS, the National Institute for Health Research or the Department of Health.
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
- Roland D, Coats T, Matheson D. Towards a conceptual framework demonstrating the effectiveness of audiovisual patient descriptions (patient video cases): a review of the current literature. BMC Med Educ. 2012;12(1):125.View ArticleGoogle Scholar
- Taylor K, Mayell A, Vandenberg S, Blanchard N, Parshuram CS. Prevalence and indications for video recording in the health care setting in North American and British paediatric hospitals. Paediatr Child Health. 2011;16(7):e57–60.Google Scholar
- Christie B, Collyer J. Do video clips add more value than audio clips? Presenting industrial research and development results using multimedia. Behav Inf Technol. 2008;27(5):395–405.View ArticleGoogle Scholar
- Schwartz D, Hartman K. It’s not television anymore: Designing digital video for learning and assessment. In: Goldman R, Derry S, Pea R, Barron B, editors. Video research in the Learning sciences. 1st ed. New Jersey: Erlbaum; 2007.Google Scholar
- Bowns I, Collins K, Walters S, McDonagh AJ. Telemedicine in dermatology: a randomised controlled trial. Health Technol Assess. 2006;10(43):58.View ArticleGoogle Scholar
- Hersh WR, Helfand M, Wallace J, Kraemer D, Patterson P, Shapiro S, et al. Clinical outcomes resulting from telemedicine interventions: a systematic review. BMC Med Inform Decis Mak. 2001;1:5.Google Scholar
- Leza R, Alesanco A, Serrano P, Ramos L, Portolés A, Aured C, et al. Analysis of Xvid Video Codec for Clinical uality Assessment in Tele-Echocardiography. Computers in Cardiology 2006;33:913–916.Google Scholar
- Webb EA, Davis L, Muir G, Lissauer T, Nanduri V, Newell SJ. Improving postgraduate clinical assessment tools: the introduction of video recordings to assess decision making. Med Teach. 2012;34(5):404–10.Google Scholar
- MacFaul R. Can real time video pictures help specialists assess acute illnessess in children? [Personal Communication]. 2005.Google Scholar
- Roland D, Wahl H, Lakhanpaul M. Blackwell N and Davies F Education by video. BMJ Careers 2011;February:1. http://careers.bmj.com/careers/advice/Education_by_video.
- Wright P, Belt S. Methods for troubleshooting a video before assessing its clinical impact. Health Informatics J. 2001;7(1):37–40.View ArticleGoogle Scholar
- Polany M. Personal knowledge:Towards a post-critical philosophy. 1st ed. Chicago: Univeristy of Chicago Press; 1958.Google Scholar
- Schon DA. The Reflective Practitioner: How Professionals Think in Action. San Francisco: Jossey Bass; 1983.Google Scholar
- Paas F, Renkl A, Sweller J. Cognitive Load Theory and Instructional Design: Recent Developments. Educ Psychol. 2003;38(1):1–4.View ArticleGoogle Scholar
- Croskerry P. Context is everything or how could I have been that stupid? Healthc Q 2009;12 Spec No Patient:e171–6.Google Scholar
- Sherbino J, Dore KL, Wood TJ, Young ME, Gaissmaier W, Kreuger S, et al. The relationship between response time and diagnostic accuracy. Acad Med. 2012;87(6):785–91.Google Scholar
- Kahneman D, Tversky A. On the study of statistical intuitions. Cognition. 1982;11(2):123–41.View ArticleGoogle Scholar
- Vranas PB. Gigerenzer’s normative critique of Kahneman and Tversky. Cognition. 2000;76(3):179–93.View ArticleGoogle Scholar
- Anderson B. Capillary refill time in adults has poor inter-observer agreement. Hong Kong J Emerg Med. 2008;15(2):71–4.Google Scholar
- Kortum P, Sullivan M. Content is king: the effect of content on the perception of video quality. Proc Human Factors Ergon Soc Annu Meet. 2004;48:1910–4.View ArticleGoogle Scholar
- Bekhof J, Reimink R, Bartels I. Eggink H and Brand PLarge observer variation of clinical assessment of dyspnoeic wheezing children. Arch Dis Child. 2015;100(7):649–53.Google Scholar