The Best Readiness Data Is the Data Nobody Had to Collect

This February, five researchers at the Army's Center for Initial Military Training published the most rigorous test yet of whether H2F works (link). They matched 28 brigades that had embedded Performance Teams against 28 that didn’t and followed both groups from FY2019 through FY2023. In the H2F brigades, the odds of a musculoskeletal injury referral fell 61% and the odds of failing the ACFT fell 22%. The authors estimated about 37,000 duty days restored per brigade per year, and a return of $8.15 for every dollar spent.

That should settle whether performance teams are worth funding. It also shows how little the Army can currently see inside them. The study could tell whether a brigade had a Performance Team. It couldn't tell how much that team was used: which soldiers saw the dietitian, whether the strength coach's programming was followed, or whether the cognitive performance specialist's calendar was full or empty. Every input was a yes or a no.

It's also a quasi-experimental study run by the program's own researchers, evaluating brigades that were deliberately hand-picked because they had the worst baselines to begin with. Skeptics will push on that selection bias, and it’s appropriate that they should. While the authors used difference-in-differences modeling to isolate trajectory changes from natural statistical regression, starting at rock bottom naturally leaves the most room for high-dollar return. The study is still incredibly valuable and points to real world impact, but the best defense against that skepticism going forward isn't just better econometrics, it’s more granular data on what practitioners and soldiers actually did, captured from day one. That’s the case I want to make here. I also want to make a harder one, about how the Army buys this capability.

It started with an embarrassment in Montreal

Australia took 180 athletes to the 1976 Montreal Olympics and came back with zero gold, one silver, and four bronze medals (link). It was the first time since 1936 the country didn't win a single gold. Even New Zealand won two, and Australia finished 32nd on the medal table. At home it was treated as a national embarrassment. The athletes described the gap bluntly. They had been working jobs and training on the side, while their competitors were full-time athletes.

The fix had been sitting on a shelf. In 1973 the government commissioned the academic John Bloomfield to write a national sports plan. Based on studies of European sports institutes, he recommended Australia build its own. A study group endorsed the idea in 1975, and then nothing happened. Montreal made it impossible to ignore. Prime Minister Malcolm Fraser's government founded the Australian Institute of Sport (link), which opened in Canberra on Australia Day 1981 (link). It started with 150 athletes across eight sports.

Sport science is what set the AIS apart. Physiology, psychology, nutrition, biomechanics and sports medicine all worked on the question of what gives an athlete an edge. The Institute built laboratories and kept refining how it assessed athletes. By 1985 the campus had a dedicated sports science and sports medicine facility. Coaches started making decisions with numbers in front of them.

The medal counts show how fast it worked (link). Australia won 5 medals in Montreal, 24 in Los Angeles in 1984, and 41 in Atlanta in 1996. In Sydney in 2000 it won 58, including 16 gold, and finished fourth on the table behind only the United States, Russia and China. Home advantage helped in Sydney, but the climb had started long before. Over the same period, the number of sports in which Australia won Olympic medals went from four to nineteen. Current or former AIS athletes won 31 of the 58 medals in Sydney.

The next step was getting measurement out of the lab. In 1999 the AIS asked Australia's Cooperative Research Centres to engineer technology that could track athletes in the field, since athletes don't perform in a controlled lab the way they do in competition. The first device was mounted on rowing boats, and later versions moved to the athlete's upper back. The technology was used only by Australian Olympic teams from 1999 to 2006, a stretch that covered the country's two most successful Games, until it was commercialized as Catapult in late 2006 (link). Fusion Sport was founded in Brisbane in 2003 to put all that athlete data in one place. Its platform, Smartabase, counted the Australian Institute of Sport among its users, and Teamworks acquired it in January 2023 (link). Smartabase is now Teamworks AMS, and the AIS is still one of our most important partners.

Australia's turnaround usually gets told as a funding story, and money mattered. What that money funded and what lasted though was the operating model: specialists from different disciplines working on the same athletes, with a record of what they did that could be checked against results. That record came mostly from the practitioners' own work, meaning testing sessions, training loads, dietary assessments and medical notes. The evidence that the system was working came out of the work itself.

Elite sport learned to measure the inputs

The rest of elite sport adopted those habits over the next two decades, and European soccer kept the best published record of what it learned. UEFA's Elite Club Injury Study has tracked injuries at Europe's top clubs for more than two decades. An 11-year analysis of 24 teams found that seasons with lower injury burden and higher player availability went with higher league finishes and more points per match (link). Healthy players win games. Nobody was surprised.

The more useful findings came when researchers asked what drives availability. Across 36 elite clubs, some medical staffs rated their communication with the head coach as poor. Those teams had higher injury burden, lower training attendance (76% vs. 83%) and lower match availability (82% vs. 88%) (link). A later analysis found that replacing the fitness coach or team doctor was associated with a jump in hamstring injury burden that season (link). Both variables came down to people: how well the staff talked to each other, and whether they stayed.

The evidence for the other pillars is uneven. A 2017 meta-analysis of 35 randomized trials found psychological skills training improved sport performance with a moderate effect, and the effect grew at follow-up. Sleep extension studies, like Stanford's work with its men's basketball team, point the same direction but are small (link). Almost all of it is associative. Clubs act on it anyway, because they can test it against their own records. Every session, meal plan and check-in sits in the same system as the injury log and the availability report. A performance director can see which inputs line up with which outcomes on their own roster.

Most military programs don't have that record.

Nutrition has the clearest link, and the military has the proof

One the clearest links between an input and availability anywhere in human performance is fueling, and the best evidence for it comes from the military.

In 2008, researchers randomized 5,201 female Navy recruits to daily calcium and vitamin D or a placebo through eight weeks of basic training (link). Stress fractures fell from 6.6% to 5.3%, a 20% reduction. It's still one of the largest nutrition trials with an injury endpoint in sport or military research. USARIEM trials at Fort Sill and Fort Jackson have since shown that the same nutrients protect bone during Army basic training, with results that depend on dose (link).

Chronic underfueling shows up in the same injuries. In a prospective study of 239 NCAA female athletes across 16 sports, athletes scored at moderate risk on the Female Athlete Triad cumulative risk assessment had 2.6 times the risk of a later bone stress injury (link). High-risk athletes had 3.8 times the risk. Half the lacrosse players in that group scored moderate or high risk, so this is not only a problem for distance runners. The same scoring approach predicted bone stress injuries in male collegiate runners (link).

Those injuries cost the Army duty days and dollars:

• Stress fractures in basic training. Across more than 580,000 Army basic trainees from 1997 to 2007, 8% of women and 1.9% of men had a stress fracture (link).

• Time lost to rehab. At Fort Jackson, trainees sent to the rehabilitation program spent an average of 62 days there, and only about half made it back to training (link). Stress fractures accounted for 39% of the injuries in that program.

• Force-wide cost. The Army Public Health Center put the 2018 cost of injuries at $4.7 billion, and $4.1 billion of that was limited-duty days (link).

• Discharges. Musculoskeletal injuries factored into nearly 70% of Army medical disability discharges between 2011 and 2016.

Fueling shows up in illness rates too. Among British Army recruits, only 21% had sufficient vitamin D in winter (link). The recruits who did had 40% lower odds of an upper respiratory infection, and each infection cost about three training days. In a randomized trial, vitamin D supplementation cut infection days by 36%. In a multi-site study of U.S. Army basic trainees, most fell well short of the military's own vitamin D intake standard (link).

Two findings from sport explain why hiring a dietitian isn't enough by itself. Stanford's Healthy Runner Project ran a dietitian-led energy availability program at two Division I programs. At one, bone stress injuries fell from 0.63 to 0.27 per athlete-year. At the other, nothing changed, and the authors pointed to differences in team culture and resources. In a separate study of elite distance runners, self-reported food diaries were a poor way to estimate energy availability, while menstrual and hormonal status flagged bone injury risk more reliably. Intake logs alone won't tell you much. Put them next to injury history, lab values and training load, and they'll show where an intervention is working and where it isn't.

That's the argument for putting a unit's dietitians, strength coaches, athletic trainers and cognitive performance specialists in one system with the soldiers they support.

Let the work produce the data

Army leaders know what happens when you ask people to feed a system that doesn't help them. In a recent Military Times piece on the H2F Management System rollout, the 199th Infantry Brigade's H2F program director explained why earlier efforts failed: “we got to get the end user buy-in to wanting to use it.” He's right, and it applies to practitioners as much as to soldiers.

The fix is to make reporting a side effect of the work. A dietitian builds a fueling plan in Teamworks Nutrition, a strength coach programs a training block, a physical therapist books a session through Hub, and a soldier logs a lift on their phone. Each of those actions is the job. Each also leaves a timestamped record of who delivered what to whom. Pull the records together and you get utilization, adherence, response time and reach, by unit and by domain, without anyone filling out a tracker on Friday afternoon.

“I need the soldier to want to use this management system so we can get the data that that organization wants to collect,” he said. “Because in the past we’ve tried it, we tried it at [the 75th Ranger] Regiment, we’ve tried it at other organizations. They failed because we got to get the end user buy-in to wanting to use it. So regardless of what comes up, what we decide to collect and what we decide to use, what metrics to use, at the end of the day, we’ve got to get that end user soldier to want to log into it.” 

-Jonathan Flynn, the H2F program director with the 199th Infantry Brigade

That's the prerequisite, and I'll state it again plainly.  This only works if practitioners actually use the tools, and they'll only use them if the tools make their day easier. A coach who has to stop coaching to feed a reporting system will do the minimum, and the data will show it. The software has to beat the clipboard and the spreadsheet on its own merits.

Then pair that engagement record with outcomes in an athlete management system, and set the baseline when the team fields. That means injury referrals, profiles, limited-duty days, ACFT scores, and lab values where you have them. A year later you can put a unit's engagement data next to its readiness trends and ask real questions. Did the companies that worked most with the dietitian see fewer bone stress injuries? Did the platoon that followed its programming hold its ACFT scores through a field-heavy quarter? That's still correlation, but it tells you far more than a yes-or-no variable. It's also how elite sport clubs decide where their next hire goes.

Tools built for the people doing the work

Elite sport and the Army aren't the same thing. What they share is that both are demanding and both are varied. Soccer, football and tennis each call for a different support philosophy. So do a Tier 1 special operations unit, a light infantry brigade and a sustainment battalion. A strength coach supporting paratroopers and a dietitian supporting mechanics on twelve-hour motor pool shifts need tools that fit their methods. One template for the whole force would fit nobody well.

Teamworks was built for that variety. It was designed by people who came out of performance departments, working alongside the practitioners who use it every day. It's now the default stack in elite athletics: more than 7,000 organizations, every team in the NFL, NHL and Premier League, and 99% of Division I athletic departments. On the warfighting side, SOCOM recently licensed Teamworks Nutrition for up to 55 dietitians supporting more than 18,000 operators (link). Our platform holds authorizations to operate with multiple service branches, in environments requiring up to IL5 (link).

You wouldn't give a performance team the software you'd give a team of accountants. You shouldn't staff one like a call center, either.

Same people, same tools

The staffing pipeline makes the comparison between sport and the military more than an analogy. H2F job postings ask for a CSCS or SCCC certification and experience with collegiate or professional athletes, the same credentials college and pro weight rooms hire on. Dietitians, athletic trainers and mental performance coaches move between sport, the services and SOF programs over their careers, and plenty move back. Many of the practitioners joining H2F teams spent years in departments that run on Teamworks. They arrive knowing the tools and the workflows. What they learn working with soldiers travels back into sport the same way.

That circulation is an asset the Army can use. It also exposes a problem with how the capability gets bought.

The labor arbitrage problem

The Army buys H2F people and H2F software on separate contracts. The people side is priced by labor category, and labor category pricing rewards one behavior: fill the seat at the lowest wage the requirement allows, then bill it at the contract rate. Posted pay for H2F strength and conditioning coaches has run roughly $50,000 to $75,000 (link). Those roles require national certification, and a master's degree at the lead level (link). Many of the coaches in those seats are excellent and could earn more elsewhere. You can attract great talent into these roles because they believe in the mission and believe that experience in a tactical environment will serve their career well, but a model that prices them as interchangeable will eventually lose them.

The soccer data shows what that turnover costs. When the fitness coach or team doctor changes, hamstring injuries go up. Every departure also takes knowledge of a unit's soldiers out the door, unless the record lives somewhere the next practitioner can find it.

A tech-enabled model runs the other way. When scheduling, communication, programming and reporting happen in one system, each practitioner spends more of the week on soldiers and less on administration. That efficiency and scalability lets a program hire more experienced people, pay them closer to what sport pays, and keep them. When the Army owns the record, it stays with the unit and follows the soldier through a PCS move, whoever holds the labor contract next year.

The middleman problem

We've seen the alternative up close. In 2013, SOCOM awarded Booz Allen Hamilton a contract worth up to $475 million for resiliency and human performance support personnel for Preservation of the Force and Family (POTFF) (link). Smartabase, now part of Teamworks, became POTFF's performance data system. It was delivered through Booz Allen as the prime integrator rather than directly to the command. In 2023, GAO reported that SOCOM hadn't completed a program evaluation of POTFF since 2016, and that delays in implementing the data system had made evaluation harder (link).

I'll own our share of that. I'll also say, from experience, that a setup where the software company can't work directly with the practitioners using its software is a bad setup. Requirements get translated twice. Fixes wait in someone else's queue. The people closest to the problem, the performance staff and the engineers who can solve it, rarely talk to each other. When SOCOM licensed Teamworks Nutrition this year, it did so directly with us because that’s the model that actually works.

The risk grows when the integrator is also a competitor. Federal acquisition rules recognize this. FAR Subpart 9.5 describes organizational conflicts of interest that arise from impaired objectivity, and from unequal access to information, including another company's proprietary data obtained while performing a contract (link). GAO has applied that reasoning to Booz Allen before. In 2016 it sustained a protest because the VA hadn't considered whether Booz Allen could objectively evaluate third-party mobile apps hosted in a cloud environment Booz Allen itself developed and ran (link). GAO didn't find misconduct, but it did find that the agency hadn't asked the question.

When one firm administers a vendor's platform, has access to its systems and its customers' data, and builds a product that competes with it for the same program, the government should ask that question every time. Better still, it should design the structure so the question never comes up.

What measurable H2F looks like

H2F is scaling toward 129 active-duty brigades, plus Guard and Reserve teams, by FY2029 (link). The Army is choosing its management system now. Four decisions in that process will determine whether the next evaluation measures engagement or only presence:

• Set baselines when each team fields, not after.

• Choose tools practitioners would pick for themselves, so the engagement record is complete.

• Keep the data under Army ownership, portable across contracts and duty stations.

• Let the people who build the software work directly with the people who use it.