Now that we have the ability to record the key health interventions in a child’s health history, we’ve begun to think how we can convey this visually. Ideally, we want a method that will allow us to quickly evaluate the status of a child’s health taking in consideration health events in the past. We hope that this will also serve as a useful tool in discussing with with CHWs cases that go wrong or right.
Here is our first attempt. The line plot on the graph is based on the child’s mid-upper arm circumference (MUAC) which provides an indicator of the child’s nutritional status and is a good proxy for a child’s overall well being. The MUAC is plotted over periods of time (P) which is intersected by health events in the child’s life. We represent each intervention: nutrition screen, CMAM treatment or malaria/diarrhea diagnosis and treatment with a line. If the intervention is performed properly ie) a malaria diagnosis and treatment with 48 hour followup – the intervention line is solid. When a followup or screening is missed a dotted line is used.
For this example, we assume that for each period P a child should receive a routine nutrition screening where a MUAC is taken. In this example the child misses his screening at P2 but receives one at P3. Between P3 and P4 the child is diagnosed with diarrhea but for some reason does not receive the proper treatment (Zinc + ORS) or perhaps misses the required 48 hour checkup. At P5, the child has a MUAC of 108 which means he has secure acute malnutrition and needs to go on plumpynut therapy. In this case, the child receives treatment through period P8 until his health rebounds. He gets malaria somewhere between P8-P9 but is properly treated and child’s nutrition status returns to a normal level.
In this chart, the history is exactly the same for the child through P7. In this case, the child misses a key nutrition screening at P8 and is not properly treated for malnutrition at P9 when is MUAC is updated. Weakened after just having recovered from malaria, the child dies, unfortunately, before P10.
Since these will also be printed out here is how the charts may look in black and white.
These are still very rough ideas and we would greatly welcome your feedback. We hope to be able to come up with something simpler / more minimal over time. If you have any other ideas or examples of how we could display this type of information please let us know.
Dear Matt and ChildCount + team,
Thank you for the opportunity to comment on your terrific project. I have been following your website for a while, although this is the first time I am commenting.
First, a basic question: Would these charts need to be legible in a very small format (i.e., on a handheld device), or do you envision them being viewed only on a computer monitor and/or in printed format? This has obvious implications for the level of visual and textual detail you might be able to consider. I have tried to tailor my comments accordingly.
Some specific points, and suggestions:
TITLE, AXES, LEGEND
In general, the more information you can convey through these headings, the less you have to detail in the graph itself.
Suggestions:
Use the title to convey the big picture. e.g., “Child Health History: nutritional status and significant health events over time.” This provides the context for the data that follow.
Label the x-axis: “Time period”. This would allow you to use a smaller, less-intrusive font for the markers P1, P2, P3, etc.
Omit the label “Legend”. Also, eliminate the faint box around the legend. Rather, leave enough white space between the graph and the legend to set it apart.
NUTRITION SCREEN #
For a quick read on a child’s health history, it seems that actual MUAC readings are not essential; the MUAC intervals on the y-axis should give a sufficient indication of where the data points lie.
Suggestion: Delete those numbers (e.g., 145, 135, 128, etc.) from the graph to ease visual clutter.
HORIZONTAL LINES (yellow / brown in colour version; dashed/dotted in b/w version)
These compete for our attention with the vertical lines. If my understanding is correct, they are there to provide important *background* information about nutritional status thresholds; they should not dominate. They also risk being misread (esp. in the b/w rendering) as “missed screening / failed treatment” events, which are also depicted using dashed lines.
Suggestion: To simplify the visual experience, use solid bands of a background colour–in muted shades–to demarcate the 3 levels now being differentiated via horizontal lines. Alternatively, you could use a single background colour (e.g., very light gray) that fills the graph space, with very fine white horizontal lines marking the health thresholds. (the white lines would be negative space, rather than additional linear elements.)
VERTICAL LINES
The purpose of the vertical lines is two-fold:
1) to identify a successful screening/treatment vs a missed screening/failed treatment; and
2) to link the health events/interventions to a time period.
As such, the lines themselves do not have to be very prominent, just legible enough to be seen as solid vs dashed. (The line weight problem is very obvious in the b/w version, where the lines become a grid that distracts from the important symbols atop each line.)
Suggestion: Use a finer line weight for both solid and dashed lines. Exception: use a heavy line weight to signify a child’s death (see comment below).
SHAPE SYMBOLS FOR HEALTH EVENTS
The use of shapes on top of vertical lines to identify particular health events or interventions is good; however, a wider range of *easily distinguishable* shapes is necessary, particularly since you have to cater to the possibility that the chart will be printed in b/w. (I think it is NOT a good idea to create a separate version for printing, ie., with added explanatory text symbols.)
Suggestion:
Do not rely on colour – i.e., don’t use a brown diamond for Plumpy Nut and a green diamond for Malaria…. it doesn’t translate into b/w, and adding an “m” for malaria only complicates things.
In addition, use larger, bolder shapes — just as the line weight should be deemphasized, the shape size should be increased.
circle = Nutritional Screening
diamond = RUTF (Plumpy Nut)
star = Malaria
square = Diarrhoea
triangle …. available if necessary for an additional health event
OUTCOME – CHILD DEATH
At first glance, the chart for Child 2 could be read as ending with a successful “screening / treatment with follow up” event — a solid line with a circle on top. The fact that the child has died is not obvious. One would not normally imagine that a MUAC would be taken as the last “event” on a child who has died. (yes, the line plot ends, and the line is black, but that is not visually arresting, given the amount of additional information on the page.)
Suggestion: Create a specific, bold marker to indicate the outcome of Child Death – i.e., a unique line style (e.g., double vertical line) that visually communicates an end-point. Identify this in the legend.
A FEW ADDITIONAL THOUGHTS….
These charts presume a fair amount of contextual knowledge — e.g., what MUAC stands for; what the lines signify at MUAC levels 110 and 125; what RUTF stands for. This is all OK, assuming these charts will be used with health workers who already know this stuff.
There are some ambiguous elements in the charts, however, that may prove problematic if you’re wanting to use the charts as an ongoing training tool with CHWs.
For example, the category “missed screening / failed treatment”… this encompasses a range of scenarios, the details of which will need to have been recorded (somewhere) in order for followup to be useful. A single “missed screening / failed treatment” event could signify any of the following scenarios:
the child was not available at the scheduled screening time;
the CHW did not show up for the screening;
the CHW was there but forgot to do the screening;
the CHW administered an incorrect treatment;
the CHW administered the correct treatment but it was not sufficient;
the CHW forgot to administer treatment.
Depending upon which scenario applied to that event, the responsibility for the problem could lie variously with the CHW, with the child’s family, or elsewhere…. info that would be crucial to learning from the experience and avoiding such a situation in the future.
Finally – I am wondering whether you envision adding additional data to these charts, such as vaccination history, or other major childhood health events (e.g., acute respiratory illness – ARI). While such additions would be great, it might get mighty complicated!
I hope these comments are helpful (albeit long).
Best wishes,
Tamsyn