A common scene if you live with a chronic illness: a clinician asks how you're doing on a scale of one to ten, you give them a number that feels honest in the moment, they nod, write something, and the conversation moves on. The most truthful thing about your week — the shape of how the bad days and the good days were arranged — never gets into the chart.
Self-tracking is the opposite. Instead of compressing weeks of variable experience into a single number on demand, you collect actual measurements on a regular cadence and look at the shape of the series. The shape is the part that's hard to fabricate and hard to forget. When you bring a curve with eight data points to your next appointment, the conversation that follows is different.
Vision is one of many things that can be tracked this way. Contrast sensitivity — how faint a pattern you can still see — is one slice of vision. Many people in MS, post-concussion, autoimmune, ME/CFS and long COVID communities already notice that something about their visual experience drifts with the rest of their symptoms, even when standard eye exams come back clean. This post is a practical guide to setting up a simple, honest tracking habit around that drift: what to measure, how often, what to log alongside it, and how to bring the result to a clinician in a form that's useful to both of you.
It's not a substitute for the people who can actually examine you. It's a way to make sure the appointment you eventually get is built on something other than your memory.
Why self-tracking works
The deepest reason self-tracking helps in chronic illness is statistical. Single-point measurements are noisy. A single result, taken on a single morning after a single night of variable sleep, captures a great deal about that morning and very little about the underlying pattern. Treated as a verdict, it's almost always misread. The same reading in context — as the eighth dot in a six-month series — is a different object. It's a sample from a distribution you can see.
The clinical literature on patient-reported outcomes has been quietly converging on the same conclusion. In a landmark trial in JAMA, Basch and colleagues randomised patients receiving chemotherapy for metastatic solid tumours to either standard care or a self-reporting arm in which patients logged twelve common symptoms regularly through a web tool. The self-reporting arm lived five months longer, used the emergency room less, and reported better quality of life (Basch, Deal, Dueck et al., 2017). The mechanism wasn't magic — the data made symptom drift visible to the care team earlier, which made earlier action possible. The lesson generalises: a continuous record beats a single snapshot, and the people taking the record are the ones living the symptoms.
A broader literature has formalised this idea under the umbrella of digital biomarkers — objective, repeatable measurements collected through everyday devices that index a physiologic or behavioural process. Coravos and colleagues' 2019 framing paper in npj Digital Medicine lays out what makes one worth taking seriously: a clear definition of what it measures, a calibration story for how it's collected, and validation against an established reference where possible (Coravos, Goldsack, Karlin et al., 2019). Not everything you measure at home meets that bar. Contrast sensitivity, measured with a calibrated tool on the same setup over time, meets enough of it to be useful as a tracking signal — not as a diagnostic test.
The practical takeaway is modest. When you bring eight monthly data points to a clinician, the conversation shifts from "how are you doing?" to "here's what's changed since last time we looked." That isn't a small shift.
What to track (vision-specific)
Before talking about contrast sensitivity specifically, it's worth naming the other vision-adjacent things that can be tracked at home. The list is longer than most people realise.
Contrast sensitivity — what we provide. Free, in your browser, three minutes for the quick mode, around seven for the full curve. Calibration runs at the start of each session so values are comparable across sessions on the same device.
Reading speed and reading endurance. Several online reading-rate tools time how many words per minute you can comfortably read at a chosen font size. Endurance — how long before your eyes ache or your accuracy drops — is the part most patients describe as actually limiting, and it's straightforward to track informally (set a timer, note the minute you started to lose the thread).
Colour perception. Free Ishihara plate sets exist online for screening colour-vision changes. Useful in particular for conditions associated with optic-nerve involvement, including MS-related optic neuritis. The home version is noisier than a clinic Ishihara — display gamut and ambient light affect it — but a trend in your own results across sessions on the same screen is informative.
Visual fatigue and screen tolerance. Time-to-symptom is the relevant measure. How many minutes of screen reading before the headache starts, the words begin to wobble, or you have to look away. Log it.
A note we'd ask you to take seriously: don't try to track everything at once. The most common failure mode in self-tracking is overshoot — too many measurements, too many variables, the habit collapses inside a month and the dataset gets abandoned. Pick one, maybe two. Contrast sensitivity and a single symptom journal is plenty. If a third measurement becomes obviously informative later, add it then.
A practical CSF tracking protocol
Here is one concrete protocol. It's simpler than the optimal one and more likely to stick.
Baseline. Take the test once in the morning, in normal indoor lighting, before any caffeine. Sit at the screen you'll use going forward. Record the device, the approximate viewing distance, and the lighting condition (e.g. "MacBook Air, ~50 cm, kitchen daylight, blinds half-open"). Save the result. The simplest way is the share-link generator on the results page: the URL itself encodes the result, so the link in your notes app is your data point.
Cadence. A reasonable default is weekly for the first month (to establish what your normal range looks like), then monthly thereafter. If you're in an active phase of your condition, more often is fine. If you're stable, less often is fine. The goal is a track, not a daily ritual.
Aligned with flares. If your condition has identifiable flare patterns, consider taking the test on a "good day" and on a "bad day" within the same week occasionally. The within-week comparison tells you how much your contrast sensitivity moves with symptoms on your particular setup, and that within-person sensitivity figure is one of the more useful things a self-tracking record can produce.
What to log alongside each reading. Three lines is enough. Time of day. Lighting (one phrase). Fatigue level on a 1–10 scale, logged at the moment of testing rather than reconstructed later. Plus any specific symptom you've noticed that day: headache, brain fog, post-exertional crash, optic-nerve discomfort, light sensitivity. The journal isn't the data; it's the context that makes the data interpretable later. Log immediately, not at end of day, because retrospective fatigue scores are systematically less accurate than in-the-moment ones — recall bias is real and it gets worse with the kind of cognitive fatigue many of the conditions this post is written for already produce.
Plot it after three or four readings. A pencil sketch on graph paper works. So does a screenshot of each result, lined up in a notes app. The thing you're looking at is the shape of the series — flat, drifting down, drifting up, oscillating around a midline, jumping after specific events — not the absolute value of any single reading. The shape is what carries information about your condition; the single values are what carry the noise.
A few honest caveats. Test-retest variation is real even in well-validated clinical instruments — the Pelli-Robson chart, the most commonly cited reference standard, has test-retest repeatability of about ±0.15 log units, and the smallest change generally considered clinically meaningful is ±0.30 log units (Pelli, Robson & Wilkins, 1988). Our test, like any consumer-screen test, sits in a noisier regime. The fix is not perfect precision; it's enough data points that the trend rises above the noise — which is what the cadence is for. The same principle that motivated efficient adaptive thresholds in psychophysics (Watson & Pelli, 1983) applies one level up, at the timescale of weeks rather than trials.
Bringing the data to your provider
Provider time is short. The fastest way to make a longitudinal record useful in a fifteen-minute appointment is to show the clinician a picture, not a spreadsheet.
A simple plot — a small chart with sessions on the x-axis and your contrast sensitivity values on the y-axis, with the journal entries available as a second sheet for context — is more readable in the first thirty seconds of a visit than any other form of the data. If you're using the share-link feature, the easiest version is one screenshot of your most recent result plus a one-line note: "Eight readings since February, here's the curve, here's what changed in March."
Suggested language to bring with the picture: "Here are my contrast sensitivity readings over the past six weeks. The shape is flat / drifting down / oscillating with my symptom days. Does this match what you'd expect for [my condition]?" That phrasing invites the clinician into the data rather than asking them to interpret a screenshot in isolation. Most clinicians, including ones unfamiliar with online contrast tests, can read a trend; what they need from you is the context (what changed, when, and what symptoms tracked the change).
The right kind of clinician depends on the situation. Neuro-optometrists — especially those listed in the Neuro-Optometric Rehabilitation Association directory — are the right audience for vision-symptom evaluation in post-concussion and many chronic-illness contexts (see our post on post-concussion vision changes for more on that). For MS patients, the audience is usually the MS neurologist or a neuro-ophthalmology referral; many MS clinics already use low-contrast Sloan letters in their standard exam, so the family of measurement is familiar territory (our MS post goes deeper on what the literature shows). For patients working through the mold / CIRS framework, the audience is the CIRS-protocol practice or your functional-medicine clinician — with the honest caveats spelled out in our CIRS post. For the general primer on what the measurement is and what 20/20 misses, see our primer.
A useful frame: you're not asking your clinician to interpret the absolute number. You're asking them to look at the shape of the line.
Pitfalls
A few honest failure modes to avoid.
Over-checking and self-induced anxiety. Don't take the test three times a day. Single-session noise will rule the results and the rising stakes you've attached to each reading will make the noise feel meaningful. Weekly is plenty for a first month; monthly is plenty after that.
Confirmation bias in calibration. If you sit down expecting bad results, you may rush calibration, sit slightly closer than usual, test in worse lighting, and produce the result you were braced for. The fix is to make the setup boringly identical each time — same device, same desk, same chair height, same time of day, same caffeine state. Same is the goal. Same is the comparison.
Recall bias in journaling. A fatigue score logged at the moment of the test is a different number from the same score reconstructed at the end of the day. Write the journal entry immediately, even one sentence, even before you look at the result. The discipline matters more than the depth.
Mistaking a number for a diagnosis. A contrast sensitivity test is a screening and tracking measurement, not a diagnostic instrument. A reduction is consistent with many things — refractive error, dry eye, fatigue, early cataract, glaucoma, MS-related changes, post-concussion changes, normal aging — and the test cannot disambiguate them. A clinician can. Don't make medication, supplement, or lifestyle changes based on a CSF reading alone.
Letting tracking replace care. The reason to collect a longitudinal record is to make appointments better, not to substitute for them. If something is changing in a way that worries you, the answer is to bring the data to a clinician sooner, not to keep collecting points until the picture is "clear." The picture is usually clearer in the room than at the spreadsheet.
The smallest version of this you can do today
Take the test once. Save the result — the share link is fine. Wait a week. Take it again. The two results, side by side, are the start of the line that all of this is about.
That's it. That's the protocol. Everything else above is refinement on the same idea. Three minutes today, three minutes next week, and you have something you didn't have before: a small piece of objective data about your own visual system, collected on your own setup, that you can bring to the people who can actually examine you. The graph is yours.
Take the test now. Come back next week.
References
- Basch, E., Deal, A. M., Dueck, A. C., Scher, H. I., Kris, M. G., Hudis, C., & Schrag, D. (2017). Overall survival results of a trial assessing patient-reported outcomes for symptom monitoring during routine cancer treatment. JAMA, 318(2), 197–198. Randomised trial showing that patients with metastatic cancer who self-reported symptoms via a web-based tool during routine chemotherapy had longer overall survival, better quality of life, and fewer emergency-department visits than usual-care controls — the load-bearing modern citation for the value of patient-collected longitudinal symptom data.
- Coravos, A., Goldsack, J. C., Karlin, D. R., Nebeker, C., Perakslis, E., Zimmerman, N., & Erb, M. K. (2019). Developing and adopting safe and effective digital biomarkers to improve patient outcomes. npj Digital Medicine, 2(1), 14. Framing paper on what qualifies as a "digital biomarker" — objective, quantifiable physiological or behavioural data collected through digital devices that indicate a clinical process — and what calibration, validation, and reporting standards make one trustworthy for clinical use.
- Pelli, D. G., Robson, J. G., & Wilkins, A. J. (1988). The design of a new letter chart for measuring contrast sensitivity. Clinical Vision Sciences, 2, 187–199. The Pelli-Robson chart paper, still the most widely used clinical letter-based contrast sensitivity test. Test-retest repeatability of about ±0.15 log units and a clinically meaningful change threshold of about ±0.30 log units — the figures any home tracking record needs to be read against.
- Watson, A. B., & Pelli, D. G. (1983). QUEST: a Bayesian adaptive psychometric method. Perception & Psychophysics, 33(2), 113–120. The foundational efficient-threshold paper underlying modern adaptive contrast-sensitivity testing. Cited here for the general principle that a precision threshold estimate needs many trials — applied here, by analogy, to the many-sessions structure of a self-tracking record.