Self vs dependent health load

A policy named for an employee can meet very different needs across the family it covers. In this claims book, cardiac categories lead the incurred cost on parent claims, infections lead on child claims, and maternity leads on spouse claims. Together, claims for someone other than the employee carry about two thirds of the cost. The family side is therefore a central part of understanding where the policy records its cost.

Laboratory results add a separate question: how the health signals differ between employees and covered family members. The answer depends on the marker and on who was tested. The large pooled blood-sugar gap becomes much smaller within most age bands, while the anemia gap remains in every age band shown.

The files cannot connect a test result to a claim. The claims file names employees, spouses, children and parents; the laboratory file records only self or others, without saying which family relationship an “other” has. They are separate datasets with no link between individual people, so the cost comparison and the test-result comparison must each stand on their own.

Two-thirds of the claims book, and two-thirds of the cost, is not the employee#

The employee is a third of the claims book. The remaining claims and incurred cost sit in the non-employee relationship groups.

Two further relationship buckets, sibling and other, are recorded but round to 0.0% on both share columns. They sit inside the combined row, which is why the four named layers come to 99.9% of claims and 100.0% of incurred once each figure is rounded.

The parent layer is where the two share columns come apart. Parents file fewer claims than employees do and carry more of the cost, because the average parent claim, ₹53,981, is the largest of the four layers. A layer whose cost arrives as bigger claims rather than as more of them behaves differently inside a policy from one that costs the same by volume: fewer claim records, each carrying a higher average. The Family Claims page carries the full claims-side breakdown, and Parent Claims goes further into the layer carrying the most incurred cost.

Claim shares count records, not people or admissions. Incurred includes open claims and is not cash paid or household spending. Claims and laboratory reports cannot be linked person by person.

The leading cost categories are not the same for each layer#

The layers do not only cost different amounts. They cost money on different things.

Each claim receives the first matching category under ordered keyword rules read across the ailment and illness fields. These are single-assignment buckets for grouping spend, not validated diagnoses, and the three categories shown for a layer lead that layer's spend rather than accounting for all of it.

Read down that table and the four layers stop looking like one population carrying more and less of the same illness. A cardiac claim and a neonatal claim are different events, and a policy meets them in different places. A benefit design built around the employee's own risk profile is built around the row that accounts for about a third of the bill. In the other two-thirds, the categories that lead are maternity, paediatric infection and respiratory illness, and the cardiac, musculoskeletal and cancer claims in the parent layer.

From the 20s through the 50s, the blood-sugar gap is under five points within each age band#

The laboratory file's headline relationship finding is a wide one. Half of the family reports carrying an HbA1c result came back elevated.

The two sets of reports do not come from similarly aged people. Among employee HbA1c reports carrying an age, 4.8% belong to someone aged 50 or over. Among family reports, 43.2% do.

Set the comparison inside each age band and the difference is much smaller.

Through the working-age bands the two columns sit within five points of each other, and in the thirties the family rate is the lower of the two by 1.4 points. Of the five bands shown, only 60 and over carries a wide within-band difference, and that row rests on 49 employee reports. Each row compares reports within the same age band; the 60+ band has no upper age limit. These comparisons do not hold exact age or sex constant, and the five rows together are not a single adjusted estimate.

The laboratory relationship field distinguishes only self from others, not spouses, children or parents. Results count reports, including repeat tests; family employment status is unknown. Age tables cover adults only.

The anemia difference does not narrow the way the blood-sugar one does#

Two other markers show a wide relationship difference in the pooled figures.

On blood sugar, a pooled difference of that size narrows to under five points from the 20s through the 50s once the comparison is made inside an age band. Anemia does not behave the same way.

The difference is wider in the 40s than in any other band shown, and narrowest of the five at 60 and over, where the employee column rests on 69 reports.

Compared separately by sex rather than by age, the two columns also stay apart, and by very different amounts. Among reports from men, anemia appears in 5.5% on the employee side (4,131 reports) and 21.7% on the family side (1,079). Among reports from women the two rates are much closer and much higher on both sides: 31.9% (2,135) against 37.2% (2,240).

Anemia varies by sex and age in this sample. Women account for 2,240 of the 3,319 family reports, compared with 2,135 of the 6,266 employee reports.

The direction is not the same on every marker#

Three other markers run the other way: the employee side of the file carries the higher share on each.

So the file does not read as one group being uniformly worse off than the other. It reads as two sets of reports with different age and sex profiles, with different patterns across the measured markers. None of these three comparisons is adjusted for either, and on vitamin D the two rates are close. The vitamin thresholds identify the report's analysis bands; they do not by themselves diagnose deficiency. The Vitamin B12 page carries that marker in full.

SGPT here is one enzyme above 40 U/L. Liver flags any of SGPT, SGOT or GGT, so its relationship figures use a different measure.

What an employer can actually do with this#

Three things this data establishes together:

About this page's data#