As of September 2013, we expected 3Q13 health services demand growth (y/y, nominal) of 3.5%, the product of 2.1% growth in unit demand and 1.4% growth in nominal pricing (Exhibit 1). Our nominal pricing forecast is up just 10 bps from last quarter’s multi-year low, while our estimate of y/y growth in unit demand is flat. Separate from our growth rate forecast, we run an independent model which handicaps the odds of a trend break in demand. This model calls for only slightly better than even odds (61%) of accelerating (3Q13 v. 2Q13) demand
The sequential dynamics in our 3Q13 estimate since our initial forecast in August are markedly negative. Pricing expectations – already muted as a direct consequence of the 2% across the board cut in Medicare provider payments that began on April 1 (Exhibit 2) – fell 10 bps month/month, primarily on further deterioration in the hospital setting and new weakness in dental services. Unit demand forecasts declined 30 bps from the August report. This was almost entirely a function of a deceleration in both total hours worked and average wage for health workers – both of which are positively correlated with unit demand
Our estimates of demand dynamics for health services rely on several measures of underlying economic activity; included among these variables are measures of general economic activity and health-system specific activity, as well as forward expectations about US economic conditions. During the federal government shutdown, the key data feeds on which our analysis relies were frozen. We have thus been unable to update estimates since our September 2013 report. Further, we normally would have expected BEA to publish actual 3Q13 demand growth results on October 30 – and even with the impasse now over (until at least February), it’s unclear how long it will be before normal BEA operations resume (data collection also stopped during the shutdown). We will update our forecast publication calendar as soon as we get clarity from the relevant federal departments about their new publication schedules
Exhibits 3, 4, and 5 provide time series of actual v. projected unit demand, price growth, and total demand, respectively
2013-2014 Flu Season
Though CDC’s weekly FluView influenza surveillance report has been on hiatus during the government shutdown, we have found alternative sources providing very comparable, rigorously backtested numbers regarding influenza-like illness (ILI) at the national level. Specifically, Google Flu Tracker uses search data algorithms to estimate flu activity in the U.S. Exhibit 6 compares Google’s and CDC’s estimates of influenza-like illness cases per 100,000 physician visits. The correlation coefficient for the two series’ is 0.88
While it is obviously far too early to make hard projections for the whole season, the first 4 weeks of the 2013-14 flu season are more similar to last year’s, than the to the average (non-pandemic), season (Exhibit 7). Estimated ILI cases per 100,000 physician visits are currently almost 30% higher than average for this time of year – but about 17% lower than this time last year. We have previously suggested that reversion to the average flu season in 2013-14 would translate into a +/- 30 bps headwind for total y/y health services demand growth during 4Q13. Adjusting for the flu activity thus far in 2013, we would very roughly estimate that if current dynamics continue, the flu effect on services unit demand growth is closer to a headwind of just 10 bps
